{
 "public_id": "AIR-2026-0075",
 "incident_id": "2026-03-snowflake-cortex-sandbox-escape",
 "date_disclosed": "2026-03-16",
 "date_occurred": null,
 "title": "Snowflake Cortex Code escapes its sandbox and executes malware: process substitution slips past the command validator",
 "class": "responsible_disclosure",
 "attack_surface": "coding_agent",
 "initial_vector": "indirect_prompt_injection_repository_content",
 "agent_capability_abused": [
  "shell_execution",
  "sub_agent_delegation",
  "sandbox_flag_configuration"
 ],
 "target_product": "Snowflake Cortex Code CLI (<= v1.0.24)",
 "target_vendor": "Snowflake",
 "autonomy_level": "copilot_with_approval",
 "impact_realized": false,
 "impact_type": [
  "remote_code_execution",
  "data_exfiltration",
  "destructive_action"
 ],
 "blast_radius": "Cortex Code CLI users processing untrusted repositories or web/database content",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI01",
   "ASI04",
   "ASI05",
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "command_validator_bypass",
   "process_substitution",
   "sandbox_escape",
   "human_in_the_loop_bypass",
   "subagent_hides_action"
  ]
 },
 "defense_present": "command validation with human approval for dangerous commands; sandbox",
 "defense_bypassed": true,
 "fix_shipped": "v1.0.25, 2026-02-28, auto-updated",
 "primary_sources": [
  "https://www.promptarmor.com/resources/snowflake-ai-escapes-sandbox-and-executes-malware"
 ],
 "verified_quote": "Because commands in process substitution expressions were not evaluated by this system, they never triggered human approval.",
 "confidence": "high",
 "notes": "PromptArmor. Cortex Code released 2026-02-02; reported 2026-02-05; fixed 2026-02-28; public 2026-03-16. Aggravating factor per the writeup: 'Cortex failed to inform the user that the command had already been run by the second-level sub-agent!' Also allowed setting a sandbox-bypass flag without user approval, enabling theft of Snowflake databases via cached auth tokens. Linked from Simon Willison's prompt-injection tag.",
 "risk_category": "security_fraud",
 "risk_secondary": [
  "privacy"
 ],
 "risk_subcategory": "Prompt Injection",
 "risk_sub_secondary": [
  "Remote Code Execution & Sandbox Escape"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "Cortex Code executed the malware past its command validator",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_repository_content",
 "guardrail_kind": [
  "human_approval",
  "sandbox_isolation"
 ],
 "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": "unsafe_actuation",
 "eai_domain": "agent_behaviour",
 "eai_secondary": [
  "privilege_escalation",
  "data_exfiltration_channels",
  "confused_deputy"
 ],
 "eai_reason": "impact rule: unsafe_actuation",
 "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"
}