{
 "id": "AIR-2026-0229",
 "slug": "cve-2026-105754-vllm-disaggregated-transport-trusts-caller-s",
 "title": "CVE-2026-105754: vLLM disaggregated transport trusts caller-supplied features",
 "summary": "The vLLM inference engine fails to properly validate caller-supplied features in its disaggregated scale-out transport path. This can lead to security bypasses within multimodal request handling pipelines.",
 "type": "attack_on_ai",
 "categories": [
  "infra_vuln"
 ],
 "severity": 2,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "vLLM"
 ],
 "incident_date": "2026-10-05",
 "incident_date_precision": "day",
 "first_reported": "2026-10-05",
 "sources": [
  {
   "url": "https://github.com/advisories/GHSA-ph72-cqr5-qpp7",
   "publisher": "GitHub Advisory Database",
   "title": "CVE-2026-105754: vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied features",
   "published_at": "2026-10-05"
  },
  {
   "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-105754",
   "publisher": "NVD",
   "title": "CVE-2026-105754: vLLM is an inference and serving engine for large language models.",
   "published_at": "2026-10-05"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_ghsa",
 "tags": [
  "CVE-2026-105754",
  "vllm",
  "inference-server"
 ],
 "created_at": "2026-10-18T06:33:31.967Z",
 "updated_at": "2026-10-18T06:33:31.967Z"
}