{
 "public_id": "AIR-2025-0068",
 "incident_id": "2025-07-lasso-identitymesh",
 "date_disclosed": "2025-07-30",
 "date_occurred": null,
 "title": "IdentityMesh: agentic lateral movement through merged identities across connected systems",
 "class": "research_demo",
 "attack_surface": "multi_agent_system",
 "initial_vector": "indirect_prompt_injection_connected_system",
 "agent_capability_abused": [
  "multi_system_connectors",
  "merged_privilege_context",
  "goal_completion_drive"
 ],
 "target_product": "MCP-based agents, AI browsers (incl. Perplexity Comet), custom crew.ai agents",
 "target_vendor": "multiple",
 "autonomy_level": "semi_autonomous",
 "impact_realized": false,
 "impact_type": [
  "lateral_movement",
  "data_exfiltration",
  "privilege_abuse"
 ],
 "blast_radius": "Enterprise assistants integrating multiple systems under a single agent identity",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI01",
   "ASI07",
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "lateral_movement",
   "identity_merge",
   "confused_deputy"
  ]
 },
 "defense_present": null,
 "defense_bypassed": true,
 "fix_shipped": null,
 "primary_sources": [
  "https://www.lasso.security/blog/identitymesh-exploiting-agentic-ai"
 ],
 "verified_quote": "The IdentityMesh vulnerability exploits a fundamental architectural weakness in Agentic AI systems: the single source of privileges problem.",
 "confidence": "high",
 "notes": "Authors Bar Lanyado, Ophir Dror, Or Oxenberg. Three attack scenarios demonstrated; PoC research, no in-the-wild exploitation claimed. Some coverage dates this to March 2026; the post itself is dated 2025-07-30.",
 "risk_category": "security_fraud",
 "risk_secondary": [
  "privacy"
 ],
 "risk_subcategory": "Prompt Injection",
 "risk_sub_secondary": [
  "Over-permissioned execution & lateral movement"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the agent moved laterally across connected systems on merged privileges",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_connected_system",
 "guardrail_kind": [
  "unknown"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI01",
 "agent_framework": "unknown",
 "model_provider": "unknown",
 "model_name": "",
 "detection_mechanism": "external_researcher",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "privilege_escalation",
 "eai_domain": "access_control_and_permissions",
 "eai_secondary": [
  "data_exfiltration_channels",
  "confused_deputy"
 ],
 "eai_reason": "impact rule: privilege_escalation",
 "asi_curated": [
  "ASI01"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI01 | added ASI07 (attack_surface=multi_agent_system + compromised exchange), ASI03 (signal=identity)",
 "submitted_by": "EAI Internal"
}