{
 "public_id": "AIR-2026-0136",
 "incident_id": "2026-06-csa-agentjacking-sentry-mcp",
 "date_disclosed": "2026-06-12",
 "date_occurred": null,
 "title": "Agentjacking: injecting instructions into Sentry error events hijacks MCP-connected coding agents",
 "class": "responsible_disclosure",
 "attack_surface": "mcp_server",
 "initial_vector": "indirect_prompt_injection_error_event",
 "agent_capability_abused": [
  "mcp_tool_use",
  "developer_identity_delegation",
  "repository_access"
 ],
 "target_product": "Sentry MCP server as consumed by Claude Code, Cursor and Codex",
 "target_vendor": "Sentry",
 "autonomy_level": "semi_autonomous",
 "impact_realized": false,
 "impact_type": [
  "agent_hijack",
  "unauthorized_action",
  "data_exfiltration"
 ],
 "blast_radius": "85% exploitation success across the three tested agents; the injection credential is a public write-only Sentry DSN",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI01",
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "indirect_prompt_injection",
   "mcp",
   "confused_deputy",
   "vendor_declined_fix"
  ]
 },
 "defense_present": "Sentry content filter added after disclosure (payload-string blocklist only)",
 "defense_bypassed": true,
 "fix_shipped": "Partial: Sentry declined root-cause remediation on 2026-06-03 and shipped only a reactive content filter",
 "primary_sources": [
  "https://labs.cloudsecurityalliance.org/research/csa-research-note-agentjacking-mcp-sentry-injection-20260612/"
 ],
 "verified_quote": "Attackers inject malicious instructions into Sentry error events using only a Sentry DSN-a public, write-only credential discoverable from browser JavaScript or GitHub search.",
 "confidence": "high",
 "notes": "CSA AI Safety Initiative research note. Sentry described the issue as 'technically not defensible' at the platform level. Quote's em-dash normalized to a hyphen. Also referenced in the wild as 'Tenet: Agentjacking'.",
 "risk_category": "security_fraud",
 "risk_secondary": [
  "privacy"
 ],
 "risk_subcategory": "Tool & Function Call Extraction / Abuse",
 "risk_sub_secondary": [
  "Prompt Injection"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the MCP-connected agents acted on injected Sentry events",
 "vector_family": "indirect_prompt_injection",
 "vector_detail": "indirect_prompt_injection_error_event",
 "guardrail_kind": [
  "permission_scope"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI01",
 "agent_framework": "unknown",
 "model_provider": "multiple",
 "model_name": "",
 "detection_mechanism": "unknown",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "unsafe_actuation",
 "eai_domain": "agent_behaviour",
 "eai_secondary": [
  "data_exfiltration_channels",
  "confused_deputy"
 ],
 "eai_reason": "impact rule: unsafe_actuation",
 "asi_curated": [
  "ASI01"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI01 | added ASI03 (signal=identity)",
 "submitted_by": "EAI Internal"
}