{
 "public_id": "AIR-2026-0180",
 "incident_id": "2026-08-google-adk-agent-to-agent-cicd",
 "date_disclosed": "2026-08-03",
 "date_occurred": null,
 "title": "I'll just call you: a low-privileged CI agent on Google's adk-python is injected into summoning a high-privileged agent across a trust boundary",
 "class": "responsible_disclosure",
 "attack_surface": "multi_agent_system",
 "initial_vector": "indirect_prompt_injection_github_issue_or_pr",
 "agent_capability_abused": [
  "agent_to_agent_triggering",
  "ci_workflow_execution",
  "git_gh_command_execution",
  "issue_and_pr_write"
 ],
 "target_product": "google/adk-python (Agent Development Kit for Python) CI/CD agents",
 "target_vendor": "Google",
 "autonomy_level": "fully_autonomous",
 "impact_realized": false,
 "impact_type": [
  "credential_theft",
  "supply_chain_compromise",
  "remote_code_execution"
 ],
 "blast_radius": "Google ADK repository and its GCP service account resources",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI02",
   "ASI01",
   "ASI04",
   "ASI07",
   "ASI05",
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "agent_to_agent_privilege_escalation",
   "ci_cd_agent",
   "allowlist_validation_gap",
   "fabricated_pr_approval"
  ]
 },
 "defense_present": "separation of low- and high-privileged workflow agents; command allowlist validation",
 "defense_bypassed": true,
 "fix_shipped": "vulnerable workflows removed, repository hardened; fix confirmed 2026-07-21 (commit 66730e9)",
 "primary_sources": [
  "https://www.pillar.security/blog/ill-just-call-you-agent-to-agent-privilege-boundary-failures-in-ci-cd-on-googles-adk-repository"
 ],
 "verified_quote": "The repository ran two classes of automated AI agents. The first class of low-privileged agents was embedded in workflows open to interaction, which were activated when a user opened a pull request or an issue.",
 "confidence": "high",
 "notes": "Pillar Security (Dan Lisichkin). Reported to Google VRP 2026-06-02 and 2026-06-05; fix confirmed 2026-07-21; published 2026-08-03. No CVE - the VRP panel reduced the reward citing social-engineering requirements. Exfiltrates ADK_TRIAGE_AGENT, GOOGLE_API_KEY and ADK_GCP_SA_KEY from the runner. Pillar's framing is the transferable part: 'The compromise lived in the _seams_, in one agent's ability to reach across a privilege boundary and set another in motion.' Compare 2025-09-cross-agent-privilege-escalation (Rehberger) on a developer workstation.",
 "risk_category": "security_fraud",
 "risk_secondary": [],
 "risk_subcategory": "Prompt Injection",
 "risk_sub_secondary": [
  "Supply Chain Compromise",
  "Credential & Secret Theft",
  "Remote Code Execution & Sandbox Escape",
  "Unauthorized Access & Account Takeover",
  "Multi-agent contagion"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the low-privileged CI agent summoned the high-privileged agent, which executed",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_github_issue_or_pr",
 "guardrail_kind": [
  "permission_scope"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI02",
 "agent_framework": "unknown",
 "model_provider": "google",
 "model_name": "",
 "detection_mechanism": "external_researcher",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "credential_theft",
 "eai_domain": "access_control_and_permissions",
 "eai_secondary": [
  "privilege_escalation",
  "supply_chain_vulnerabilities",
  "confused_deputy"
 ],
 "eai_reason": "impact rule: credential_theft",
 "asi_curated": [
  "ASI02",
  "ASI01"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI02,ASI01 | added ASI04 (vector=supply_chain), ASI07 (vector=inter_agent), ASI05 (impact=code_exec), ASI03 (eai_category=credential_theft)",
 "submitted_by": "EAI Internal"
}