{
 "id": "AIR-2025-0023",
 "slug": "cve-2025-6051-redos-vulnerability-in-hugging-face-transforme",
 "title": "CVE-2025-6051: ReDoS vulnerability in Hugging Face Transformers EnglishNormalizer",
 "summary": "A denial of service vulnerability was discovered in the Hugging Face Transformers library affecting EnglishNormalizer. The flaw allows crafted inputs to cause excessive CPU consumption, which has been patched in version 4.53.0.",
 "type": "attack_on_ai",
 "categories": [
  "infra_vuln"
 ],
 "severity": 2,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "Transformers"
 ],
 "incident_date": "2025-09-14",
 "incident_date_precision": "day",
 "first_reported": "2025-09-14",
 "sources": [
  {
   "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-6051",
   "publisher": "NVD",
   "title": "CVE-2025-6051: A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library,…",
   "published_at": "2025-09-14"
  },
  {
   "url": "https://github.com/advisories/GHSA-rcv9-qm8p-9p6j",
   "publisher": "GitHub Advisory Database",
   "title": "CVE-2025-6051: Hugging Face Transformers library has Regular Expression Denial of Service",
   "published_at": "2025-09-14"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_nvd",
 "tags": [
  "CVE-2025-6051",
  "redos",
  "huggingface"
 ],
 "created_at": "2026-10-18T06:33:31.967Z",
 "updated_at": "2026-10-18T07:00:14.101Z"
}