{
 "id": "AIR-2026-0770",
 "slug": "cve-2026-54236-vllm-incomplete-fix-leaks-pil-repr-addresses",
 "title": "CVE-2026-54236: vLLM Incomplete Fix Leaks PIL Repr Addresses",
 "summary": "An incomplete fix for a prior vulnerability in vLLM leads to information disclosure of memory addresses via the Anthropic API router. The flaw has been documented with a vendor advisory and patch.",
 "type": "attack_on_ai",
 "categories": [
  "infra_vuln"
 ],
 "severity": 2,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "vLLM"
 ],
 "incident_date": "2026-06-17",
 "incident_date_precision": "day",
 "first_reported": "2026-06-17",
 "sources": [
  {
   "url": "https://github.com/advisories/GHSA-hgg8-fqqc-vfmw",
   "publisher": "GitHub Advisory Database",
   "title": "CVE-2026-54236: vLLM: incomplete CVE-2026-22778 fix leaks PIL repr addresses via Anthropic router",
   "published_at": "2026-06-17"
  },
  {
   "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-54236",
   "publisher": "NVD",
   "title": "CVE-2026-54236: vLLM is an inference and serving engine for large language models (LLMs).",
   "published_at": "2026-06-22"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_ghsa",
 "tags": [
  "CVE-2026-54236",
  "vllm",
  "information-disclosure"
 ],
 "created_at": "2026-10-18T07:00:14.101Z",
 "updated_at": "2026-10-18T18:20:34.882Z"
}