{
 "id": "AIR-2026-0437",
 "slug": "cve-2026-22807-vllm-rce-via-auto-map-dynamic-module-loading",
 "title": "CVE-2026-22807: vLLM RCE via auto_map Dynamic Module Loading",
 "summary": "vLLM loads Hugging Face auto_map dynamic modules without gating on trust_remote_code, allowing arbitrary code execution at server startup. The vulnerability can be exploited by influencing the model repository path.",
 "type": "attack_on_ai",
 "categories": [
  "infra_vuln"
 ],
 "severity": 2,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "vLLM"
 ],
 "incident_date": "2026-01-21",
 "incident_date_precision": "day",
 "first_reported": "2026-01-21",
 "sources": [
  {
   "url": "https://github.com/advisories/GHSA-2pc9-4j83-qjmr",
   "publisher": "GitHub Advisory Database",
   "title": "CVE-2026-22807: vLLM affected by RCE via auto_map dynamic module loading during model initialization",
   "published_at": "2026-01-21"
  },
  {
   "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-22807",
   "publisher": "NVD",
   "title": "CVE-2026-22807: vLLM is an inference and serving engine for large language models (LLMs).",
   "published_at": "2026-01-21"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_ghsa",
 "tags": [
  "CVE-2026-22807",
  "rce",
  "infra_vuln"
 ],
 "created_at": "2026-10-18T07:00:14.101Z",
 "updated_at": "2026-10-18T10:04:21.885Z"
}
