{"dataType": "CVE_RECORD", "cveMetadata": {"state": "PUBLISHED", "cveId": "CVE-2026-31253", "assignerOrgId": "8254265b-2729-46b6-b9e3-3dfca2d5bfca", "assignerShortName": "mitre", "dateUpdated": "2026-05-11T15:40:10.644Z", "dateReserved": "2026-03-09T00:00:00.000Z", "datePublished": "2026-05-11T00:00:00.000Z"}, "containers": {"cna": {"providerMetadata": {"orgId": "8254265b-2729-46b6-b9e3-3dfca2d5bfca", "shortName": "mitre", "dateUpdated": "2026-05-11T15:40:10.644Z"}, "descriptions": [{"lang": "en", "value": "The flash-attention training framework thru commit e724e2588cbe754beb97cf7c011b5e7e34119e62 (2025-13-04) contains an insecure deserialization vulnerability (CWE-502) in its checkpoint loading mechanism. The load_checkpoint() function in checkpoint.py and the checkpoint loading code in eval.py use torch.load() without enabling the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by providing a maliciously crafted checkpoint file. When a victim loads this checkpoint during model warmstarting or evaluation, arbitrary code is executed on the victim's system."}], "affected": [{"vendor": "n/a", "product": "n/a", "versions": [{"version": "n/a", "status": "affected"}]}], "references": [{"url": "https://github.com/Dao-AILab/flash-attention"}, {"url": "https://www.notion.so/CVE-2026-31253-35d1e1393188813f9e77e2038104bc49"}], "problemTypes": [{"descriptions": [{"type": "text", "lang": "en", "description": "n/a"}]}]}}, "dataVersion": "5.2"}

{}

{"cveTags": [], "descriptions": [{"lang": "en", "value": "The flash-attention training framework thru commit e724e2588cbe754beb97cf7c011b5e7e34119e62 (2025-13-04) contains an insecure deserialization vulnerability (CWE-502) in its checkpoint loading mechanism. The load_checkpoint() function in checkpoint.py and the checkpoint loading code in eval.py use torch.load() without enabling the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by providing a maliciously crafted checkpoint file. When a victim loads this checkpoint during model warmstarting or evaluation, arbitrary code is executed on the victim's system."}], "id": "CVE-2026-31253", "lastModified": "2026-05-11T17:16:20.307", "metrics": {}, "published": "2026-05-11T17:16:20.307", "references": [{"source": "cve@mitre.org", "url": "https://github.com/Dao-AILab/flash-attention"}, {"source": "cve@mitre.org", "url": "https://www.notion.so/CVE-2026-31253-35d1e1393188813f9e77e2038104bc49"}], "sourceIdentifier": "cve@mitre.org", "vulnStatus": "Received"}

{}

{"created": "2026-05-11T17:45:26.342359+00:00", "title": "Insecure Deserialization in flash\u2011attention Checkpoint Loading Allows Arbitrary Code Execution", "updated": "2026-05-11T17:45:26.342374+00:00", "vendors": [], "weaknesses": ["CWE-502"]}