{
  "id": "anthropic-novice-malware-builder-2025-03",
  "name": "Novice actor LLM-enabled malware builder progression",
  "date_reported": "2025-04-23",
  "researcher": "Anthropic Threat Intelligence",
  "org": "Anthropic",
  "sources": [
    {
      "url": "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
    }
  ],
  "magicsword_coverage": [],
  "borderline": true,
  "borderline_note": "Capability elevation case; Anthropic did not confirm real-world deployment.",
  "ai_role": "llm_generated_malware",
  "ai_role_secondary": null,
  "ai_role_detail": "LLM-assisted malware development (skill amplification); deployment unconfirmed",
  "technique_summary": "Limited-skill actor used Claude to evolve from basic batch scripts to GUI malware builder focused on evasion and persistence, plus doxing/remote-access tooling with facial recognition and dark-web scanning features.",
  "initial_access": "unknown",
  "ai_tooling_and_models": [
    "Claude"
  ],
  "other_tooling": [],
  "victims_and_sectors": "unknown",
  "regions": [
    "unknown"
  ],
  "attribution_or_cluster": "unknown",
  "scale_and_metrics": "unknown",
  "prevention_angle": "Model abuse detection; classic malware detections still apply to outputs",
  "related_magicsword_projects": [
    "LOLBAS"
  ],
  "source_notes": null,
  "last_reviewed": "2026-10-20",
  "attack_flow": {
    "nodes": [
      {
        "id": "step-1",
        "title": "Limited-skill actor requests help",
        "description": "The actor starts with basic batch-script development.",
        "kind": "access",
        "source_urls": [
          "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
        ]
      },
      {
        "id": "step-2",
        "title": "Claude assists iteration",
        "description": "Model assistance supports increasingly complex malware tooling.",
        "kind": "ai",
        "source_urls": [
          "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
        ]
      },
      {
        "id": "step-3",
        "title": "GUI builder takes shape",
        "description": "The reported builder adds evasion, persistence and remote-access features.",
        "kind": "execution",
        "source_urls": [
          "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
        ]
      },
      {
        "id": "step-4",
        "title": "Deployment is not established",
        "description": "The entry describes tool development rather than a confirmed victim intrusion.",
        "kind": "unknown",
        "source_urls": [
          "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
        ]
      }
    ],
    "edges": [
      {
        "source": "step-1",
        "target": "step-2"
      },
      {
        "source": "step-2",
        "target": "step-3"
      },
      {
        "source": "step-3",
        "target": "step-4"
      }
    ]
  },
  "atlas_techniques": [
    {
      "id": "AML.T0016.002",
      "rationale": "The supplied source explicitly describes generative-model tooling assisting the reported activity; this mapping does not establish additional autonomy or victim impact.",
      "source_urls": [
        "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
      ]
    }
  ],
  "category": "adversary-use-of-ai",
  "ai_involvement_confidence": "unverified",
  "detections": [
    {
      "name": "Review attack-specific indicators",
      "format": "guidance",
      "notes": "Monitor for the techniques described in Novice actor LLM-enabled malware builder progression. Refer to sources for specific IOCs and TTPs."
    }
  ]
}