{
 "id": "AIR-2025-0286",
 "slug": "cve-2025-48944-vllm-tool-schema-allows-dos-via-malformed-fie",
 "title": "CVE-2025-48944: vLLM Tool Schema Allows DoS via Malformed Fields",
 "summary": "The vLLM backend endpoint lacks input validation for pattern and type fields during tool invocations. Malformed requests can crash inference workers, causing service downtime until manual restarts occur.",
 "type": "attack_on_ai",
 "categories": [
  "infra_vuln"
 ],
 "severity": 2,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "vLLM"
 ],
 "incident_date": "2025-05-28",
 "incident_date_precision": "day",
 "first_reported": "2025-05-28",
 "sources": [
  {
   "url": "https://github.com/advisories/GHSA-vrq3-r879-7m65",
   "publisher": "GitHub Advisory Database",
   "title": "CVE-2025-48944: vLLM Tool Schema allows DoS via Malformed pattern and type Fields",
   "published_at": "2025-05-28"
  },
  {
   "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-48944",
   "publisher": "NVD",
   "title": "CVE-2025-48944: vLLM is an inference and serving engine for large language models (LLMs).",
   "published_at": "2025-05-30"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_ghsa",
 "tags": [
  "CVE-2025-48944",
  "dos",
  "vllm"
 ],
 "created_at": "2026-10-18T07:00:14.101Z",
 "updated_at": "2026-10-18T10:04:21.885Z"
}