{
 "public_id": "AIR-2026-0147",
 "incident_id": "2026-07-promptarmor-agentic-auto-review-approves-malware",
 "date_disclosed": "2026-07",
 "date_occurred": null,
 "title": "Agentic auto-review approves malware: agent-in-the-loop approval modes in Codex and Claude Code green-light a malicious npm install from an HTML-comment injection",
 "class": "research_demo",
 "attack_surface": "coding_agent",
 "initial_vector": "indirect_prompt_injection_github_issue_html_comment",
 "agent_capability_abused": [
  "auto_approval_by_model",
  "package_installation",
  "post_install_script_execution"
 ],
 "target_product": "OpenAI Codex; Anthropic Claude Code (auto/approve-for-me modes)",
 "target_vendor": "OpenAI / Anthropic",
 "autonomy_level": "semi_autonomous",
 "impact_realized": false,
 "impact_type": [
  "remote_code_execution",
  "malware_execution"
 ],
 "blast_radius": "orgs running coding agents with model-delegated command approval",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI01",
   "ASI04",
   "ASI05",
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "approval_delegated_to_model",
   "supply_chain",
   "postinstall_script",
   "hidden_html_comment"
  ]
 },
 "defense_present": "command approval - but delegated to the model rather than a human",
 "defense_bypassed": true,
 "fix_shipped": "mitigation via managed settings: Claude permissions.disableAutoMode='disable'; Codex allowed_approval_reviewers=['user']",
 "primary_sources": [
  "https://www.promptarmor.com/resources/agentic-auto-review-approves-malware"
 ],
 "verified_quote": "A post-install script in the NPM package runs immediately upon installation and executes with the user's full privileges.",
 "confidence": "high",
 "notes": "PROVENANCE CAVEAT: no publication date on the page; 2026-07 in the incident_id is provisional and date_disclosed is left null. PromptArmor. Complements 2026-08-anthropic-claude-code-automode-rce (Rehberger) - both attack the premise that a classifier model can safely stand in for the human approver. | date_disclosed is recorded to the month only; no source establishes the exact day.",
 "risk_category": "security_fraud",
 "risk_secondary": [],
 "risk_subcategory": "Prompt Injection",
 "risk_sub_secondary": [
  "Supply Chain Compromise",
  "Remote Code Execution & Sandbox Escape"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the model auto-approved and the agent installed the malicious npm package",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_github_issue_html_comment",
 "guardrail_kind": [
  "human_approval"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI01",
 "agent_framework": "unknown",
 "model_provider": "multiple",
 "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": [
  "policy_drift",
  "confused_deputy",
  "supply_chain_vulnerabilities"
 ],
 "eai_reason": "impact rule: privilege_escalation",
 "asi_curated": [
  "ASI01",
  "ASI04"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI01,ASI04 | added ASI05 (impact=code_exec), ASI03 (signal=identity)",
 "submitted_by": "EAI Internal"
}