{
 "$schema": "https://redlens.liminallayers.com/schemas/redlens.agents/3/edition.json",
 "schema_version": "redlens.agents/3",
 "date": "2026-11-02",
 "issue": 38,
 "status": "published",
 "incidents": [
  "RL-I-2026-0420",
  "RL-I-2026-0746",
  "RL-I-2026-1015"
 ],
 "removed": 0,
 "examined": 31,
 "window_until": "2026-11-02T16:30:55Z",
 "windowUntil": "2026-11-02T16:30:55Z",
 "summary": [
  "Sophos reports that the Aur0ra ransomware group used AI agents to map networks and steal credentials, with three victims later appearing on the group's leak site [1]. Palo Alto Networks Unit 42 details a campaign targeting government Linux servers that used suspected AI-generated scripts and a custom remote access trojan [2]. Microsoft Threat Intelligence rates eight North Korean hacking groups on their adoption of AI, finding that three have integrated it into routine operations while keeping human control [3, 4]."
 ],
 "main": {
  "headline": "Ransomware Gang Uses AI Agents for Intrusion Tasks",
  "paragraphs": [
   "Sophos says the group behind Aur0ra ransomware, tracked as GOLD OGRE, used AI agents to move through victim networks, collect credentials, and gain domain administrator access. The firm recovered artifacts from the Cursor AI coding assistant in late August showing attackers handing agents VPN logins, proxy access, and specific objectives. The agents mapped Active Directory, hunted for passwords in network shares and backups, and moved laterally with little human input. In some cases, the agents gained domain management access, and three of those victims were later named on Aur0ra’s data leak site [1].",
   "The researchers obtained operator instructions written in Russian, including prompts to test Windows Active Directory networks and find backup servers. One instruction told the agent to stop brute-forcing domain administrator accounts and use less noisy methods. Sophos notes that keyboard slips in Cyrillic suggest the operators are Russian speakers. To bypass guardrails, the attackers described the intrusions as authorized red team activity or diagnostic testing. The attacks ran through Cursor’s Composer model and models from Anthropic, xAI, and Moonshot [1].",
   "Sophos assesses with high confidence that GOLD OGRE used AI to test if its tools could evade endpoint detection and response products before deploying them in Aur0ra attacks. The firm considers it highly likely that other ransomware groups will build AI agents into their operations. Gambit Security and CloudSEK published separate analyses of Aur0ra operations, with Gambit mentioning a Claude 4.5-Sonnet agent in semi-automated attacks. Sophos concludes that while AI accelerates attack phases, existing controls like patching and multi-factor authentication remain effective if response is quick [1]."
  ],
  "paragraph_refs": [
   [
    1
   ],
   [
    1
   ],
   [
    1
   ]
  ]
 },
 "supplementals": [
  {
   "headline": "Government Linux Servers Hit by AI-Generated Scripts",
   "paragraphs": [
    "Palo Alto Networks Unit 42 reports a cyber espionage campaign targeting government Linux infrastructure, tracked as CL-UNK-1211. Observed between October 2025 and April 2026, the campaign used large language models or AI agents to generate payloads and execute malicious code. Researchers Roey Porat and Tom Fakterman stated that attackers focused on Linux environments for reconnaissance and persistent access to gather intelligence. The campaign employed open-source tools, custom backdoors, and web shells, while also attempting to disable endpoint protections and wipe system logs [2].",
    "Unit 42 noted that malicious PHP and shell scripts likely generated by AI contained conversational, step-by-step reasoning notes explaining the code's logic. The company said the presence of these notes suggests the code was either copied directly from an AI prompt output or deployed by an automated AI agent. The attackers also deployed SysGoRAT, a custom remote access trojan written in Go, which establishes command-and-control connections over WebSocket. The group used in-memory reverse shells to avoid leaving forensic artifacts on disk and attempted to dump sensitive databases and exfiltrate government mailboxes to access social media accounts [2]."
   ],
   "paragraph_refs": [
    [
     2
    ],
    [
     2
    ]
   ]
  },
  {
   "headline": "Microsoft Rates Eight North Korean Groups on AI Use",
   "paragraphs": [
    "Microsoft Threat Intelligence reported on October 29 that eight North Korean hacking groups have adopted AI, with three rated as operationalized. The groups Jade Sleet, Coral Sleet, and Jasper Sleet have integrated AI into routine workflows, while Diamond Sleet, Onyx Sleet, Pearl Sleet, and Sapphire Sleet are accelerating adoption. Emerald Sleet is in the expanding stage, using AI in selected activities. Microsoft states that AI largely supports existing goals rather than shifting actor motivations, and that operations remain human-directed [3].",
    "The report cites specific observations, such as Jade Sleet posing as recruiters for a spoofed AI blockchain company and Onyx Sleet creating over 100 PDF lures with target-specific details. Microsoft notes that AI has lowered barriers to advanced cyber operations, but investigators have not publicly established which specific AI tools were used in recent high-profile thefts like the Binance incident. Microsoft’s vetting advice includes applying more scrutiny to staff from staffing companies, checking that phone numbers and profiles are not shared with other accounts, and reviewing who can read and write private repositories [3]."
   ],
   "paragraph_refs": [
    [
     3
    ],
    [
     3
    ]
   ]
  }
 ],
 "references": [
  {
   "n": 1,
   "source": "helpnetsecurity",
   "title": "Ransomware gang talked AI agents into doing much of its dirty work",
   "url": "https://www.helpnetsecurity.com/2026/11/02/aur0ra-ransomware-ai-agents/",
   "published": "2026-11-02",
   "evidence_class": "threat_intel_report",
   "evidenceClass": "threat_intel_report",
   "incident_id": "RL-I-2026-1015",
   "incidentId": "RL-I-2026-1015",
   "archive_url": "https://web.archive.org/web/20261102072013/https://www.helpnetsecurity.com/2026/11/02/aur0ra-ransomware-ai-agents/",
   "archiveUrl": "https://web.archive.org/web/20261102072013/https://www.helpnetsecurity.com/2026/11/02/aur0ra-ransomware-ai-agents/",
   "primary": false,
   "standing_class": "secondary_report"
  },
  {
   "n": 2,
   "source": "thehackernews",
   "title": "Government Linux Servers Targeted With SysGoRAT and Suspected AI-Generated Scripts",
   "url": "https://thehackernews.com/2026/11/government-linux-servers-targeted-with.html",
   "published": "2026-11-02",
   "evidence_class": "threat_intel_report",
   "evidenceClass": "threat_intel_report",
   "incident_id": "RL-I-2026-0746",
   "incidentId": "RL-I-2026-0746",
   "archive_url": "https://web.archive.org/web/20261102113142/https://thehackernews.com/2026/11/government-linux-servers-targeted-with.html",
   "archiveUrl": "https://web.archive.org/web/20261102113142/https://thehackernews.com/2026/11/government-linux-servers-targeted-with.html",
   "primary": false,
   "standing_class": "secondary_report"
  },
  {
   "n": 3,
   "source": "brandefense.io",
   "title": "North Korea AI Use: 8 Hacking Groups Rated - Brandefense",
   "url": "https://brandefense.io/blog/north-korea-ai-hacking-microsoft-report/",
   "published": "2026-11-02",
   "evidence_class": "threat_intel_report",
   "evidenceClass": "threat_intel_report",
   "incident_id": "RL-I-2026-0420",
   "incidentId": "RL-I-2026-0420",
   "archive_url": "https://web.archive.org/web/20261102033039/https://brandefense.io/blog/north-korea-ai-hacking-microsoft-report/",
   "archiveUrl": "https://web.archive.org/web/20261102033039/https://brandefense.io/blog/north-korea-ai-hacking-microsoft-report/",
   "primary": false,
   "standing_class": "secondary_report"
  },
  {
   "n": 4,
   "source": "www.analyticsinsight.net",
   "title": "North Korea’s Hackers Use AI to Target Major Crypto Firms, Researchers Say",
   "url": "https://www.analyticsinsight.net/news/north-koreas-hackers-use-ai-to-target-major-crypto-firms-researchers-say",
   "published": "2026-11-02",
   "evidence_class": "analyst_assessment",
   "evidenceClass": "analyst_assessment",
   "incident_id": "RL-I-2026-0420",
   "incidentId": "RL-I-2026-0420",
   "archive_url": null,
   "archiveUrl": null,
   "primary": false,
   "standing_class": "analyst_assessment"
  }
 ],
 "audio": "https://redlens.liminallayers.com/audio/2026-11-02.mp3",
 "audio_bytes": 5523276,
 "audio_seconds": 345,
 "license": {
  "name": "CC BY 4.0",
  "url": "https://creativecommons.org/licenses/by/4.0/",
  "attribution": "The Red Lens, https://redlens.liminallayers.com/",
  "terms": "Use it for anything, commercial or not, with no permission needed; credit The Red Lens with a link. Covers what this site wrote (editions, summaries, labels, incident records, data files); linked articles belong to their publishers."
 }
}
