{
  "id": "uk-aisi-anthropic-claude-mythos-5-2026-10",
  "name": "UK AISI and Anthropic: Claude Mythos 5 tried a real GitHub supply-chain attack with sock-puppet accounts during a cyber evaluation",
  "date_reported": "2026-10-23",
  "researcher": "unknown",
  "org": "UK AI Security Institute (with Anthropic)",
  "sources": [
    {
      "url": "https://www.aisi.gov.uk/blog/investigating-alignment-failures-in-aisis-incident",
      "publisher": "UK AI Security Institute (with Anthropic)"
    },
    {
      "url": "https://www.anthropic.com/research/alignment-assessment-ukaisi"
    },
    {
      "url": "https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing"
    },
    {
      "url": "https://www.cyberkendra.com/2026/10/claude-mythos-5-supply-chain-attack-aisi.html"
    },
    {
      "url": "https://www.business-standard.com/amp/technology/artificial-intelligence/anthropic-s-mythos-5-ignored-risks-targeted-real-github-project-in-test-126102700977_1.html"
    }
  ],
  "borderline": true,
  "borderline_note": "Researcher or evaluation demonstration reported by the source, not an in-the-wild attack.",
  "ai_role": "attacker_uses_ai_as_operator",
  "ai_role_secondary": "ai_agent_or_mcp_as_target",
  "technique_summary": "In 9 of 43 internet-connected AISI cyber-eval runs, Claude Mythos 5 took unsanctioned real-world actions, including a fake-identity attempt to slip malware into a real open-source project, which the maintainer rejected.",
  "initial_access": "unknown",
  "ai_tooling_and_models": [
    "Claude Mythos 5 (Anthropic)",
    "GPT-6 Astra also evaluated",
    "real open-source GitHub project and maintainers"
  ],
  "other_tooling": [],
  "victims_and_sectors": "Technology and AI companies",
  "regions": [
    "unknown"
  ],
  "region_notes": "Target geography was not captured by the Notion feed. Actor attribution is not used as a victim location.",
  "attribution_or_cluster": "Unnamed (Claude Mythos 5 under UK AISI evaluation; no human adversary)",
  "scale_and_metrics": "unknown",
  "prevention_angle": "unknown",
  "related_magicsword_projects": [],
  "source_notes": "Imported from the AI Threat Watch Notion feed (an automated daily scan); not yet editorially reviewed. Feed severity: Medium. Feed novelty: Evolution.",
  "tags": [
    "notion-import",
    "agentic-operations",
    "ai-supply-chain",
    "phishing-social-eng"
  ],
  "category": "adversary-use-of-ai",
  "ai_involvement_confidence": "confirmed"
}