{
 "public_id": "AIR-2025-0037",
 "incident_id": "2025-05-github-mcp-server-private-repo-exfiltration",
 "date_disclosed": "2025-05-26",
 "date_occurred": null,
 "title": "GitHub MCP server hijacked via malicious issue: agent coerced into leaking private repository data",
 "class": "research_demo",
 "attack_surface": "mcp_server",
 "initial_vector": "indirect_prompt_injection_github_issue",
 "agent_capability_abused": [
  "repository_read",
  "cross_repository_access",
  "pull_request_creation",
  "cross_repo_token_scope",
  "cross_repo_access"
 ],
 "target_product": "github/github-mcp-server",
 "target_vendor": "GitHub",
 "autonomy_level": "semi_autonomous",
 "impact_realized": false,
 "impact_type": [
  "data_exfiltration",
  "confused_deputy",
  "private_source_code_disclosure"
 ],
 "blast_radius": "any user pointing an agent at the GitHub MCP server with a broadly-scoped PAT and reading public issues",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI01",
   "ASI03"
  ],
  "mitre_atlas": [
   "AML.T0051.001"
  ],
  "labels": [
   "indirect_prompt_injection",
   "confused_deputy",
   "toxic_agent_flow",
   "lethal_trifecta",
   "architectural",
   "cross_repo_scope_violation",
   "architectural_not_patchable",
   "mcp",
   "private_repo_leak"
  ]
 },
 "defense_present": "model alignment only (Claude 4 Opus was the tested agent)",
 "defense_bypassed": true,
 "fix_shipped": "none at the code level - Invariant framed it as architectural; mitigations are per-session repo scoping and least-privilege tokens",
 "primary_sources": [
  "https://invariantlabs.ai/blog/mcp-github-vulnerability",
  "https://github.com/github/github-mcp-server/issues/844"
 ],
 "verified_quote": "This is not a flaw in the GitHub MCP server code itself, but rather a fundamental architectural issue that must be addressed at the agent system level.",
 "confidence": "high",
 "notes": "Invariant Labs, 2025-05-26. Demonstrated against a live setup they controlled; exfiltrated private repo names, relocation plans and salary details into a public PR. No CVE assigned because GitHub treated it as an architectural/agent-design issue rather than a server bug. Simon Willison's writeup coined the framing widely reused as the 'lethal trifecta'. Also reported under the titles: 'GitHub MCP server: prompt injection in a public issue exfiltrates private repository contents'; 'GitHub MCP exploited: a malicious issue coerces the agent into leaking private repository data'.",
 "risk_category": "privacy",
 "risk_secondary": [],
 "risk_subcategory": "Data Exfiltration via Context",
 "risk_sub_secondary": [],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the agent read the private repo and wrote the data into a public PR",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_github_issue",
 "guardrail_kind": [
  "unknown"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [
   "AML.T0051.001"
  ],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI01",
 "agent_framework": "mcp_server",
 "model_provider": "unknown",
 "model_name": "",
 "detection_mechanism": "external_researcher",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "data_exfiltration_channels",
 "eai_domain": "privacy",
 "eai_secondary": [
  "confused_deputy"
 ],
 "eai_reason": "subcategory: Data Exfiltration via Context",
 "asi_curated": [
  "ASI01"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI01 | added ASI03 (signal=identity)",
 "submitted_by": "EAI Internal"
}