{
 "public_id": "AIR-2026-0003",
 "incident_id": "2026-01-sentinelone-claude-code-dependency-hijack",
 "date_disclosed": "2026-01-06",
 "date_occurred": null,
 "title": "Marketplace skill in Claude Code silently redirects dependency installs to an attacker-controlled source",
 "class": "research_demo",
 "attack_surface": "skill_plugin",
 "initial_vector": "malicious_skill_from_unofficial_marketplace",
 "agent_capability_abused": [
  "dependency_installation",
  "persistent_skill_context"
 ],
 "target_product": "Claude Code plugin/skill marketplaces",
 "target_vendor": "Anthropic (Claude Code) / third-party marketplaces",
 "autonomy_level": "copilot_with_approval",
 "impact_realized": false,
 "impact_type": [
  "malware_distribution",
  "credential_theft"
 ],
 "blast_radius": "any project the agent installs dependencies into after the skill is enabled",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI05",
   "ASI04",
   "ASI02"
  ],
  "mitre_atlas": [],
  "labels": [
   "dependency_hijack",
   "skill_persistence",
   "index_url_redirection",
   "trust_boundary_expansion"
  ]
 },
 "defense_present": "none - skills persist across sessions once enabled",
 "defense_bypassed": true,
 "fix_shipped": "n/a (proof of concept)",
 "primary_sources": [
  "https://www.sentinelone.com/blog/marketplace-skills-and-dependency-hijack-in-claude-code/"
 ],
 "verified_quote": "If your coding assistant can fetch and execute code on your behalf, every plugin installed joins your trust boundary",
 "confidence": "high",
 "notes": "SentinelOne proof of concept, not observed in the wild. A benign-looking plugin exposes a dependency-management skill; when the developer asks the agent to install a common Python library the skill quietly redirects to an attacker-controlled index, so a trojanised build lands in the project while imports and example code still work.",
 "risk_category": "security_fraud",
 "risk_secondary": [],
 "risk_subcategory": "Supply Chain Compromise",
 "risk_sub_secondary": [
  "Credential & Secret Theft",
  "Persistence & Covert C2"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the skill redirected the agent's dependency installs to the attacker",
 "vector_family": "malicious_component",
 "vector_detail": "malicious_skill_from_unofficial_marketplace",
 "guardrail_kind": [
  "none"
 ],
 "guardrail_outcome": "not_applicable",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI05",
 "agent_framework": "claude_code",
 "model_provider": "anthropic",
 "model_name": "",
 "detection_mechanism": "external_researcher",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "supply_chain_vulnerabilities",
 "eai_domain": "tool_misuse",
 "eai_secondary": [
  "credential_theft"
 ],
 "eai_reason": "subcategory: Supply Chain Compromise",
 "asi_curated": [
  "ASI05"
 ],
 "asi_scope": "in_scope",
 "asi_provenance": "curated_plus_rule",
 "asi_reason": "curated: ASI05 | added ASI04 (vector_family=malicious_component), ASI02 (impact=tool_misuse)",
 "submitted_by": "EAI Internal"
}