NIST deprecates RSA-2048 from 2030. Find every one of them before attackers do.
🐍
🐍 PyPI
Not in CISA KEV
HIGH severity

CVE-2026-31253 — flash-attn

HIGH

CVE-2026-31253 is a high-severity (CVSS 7.3) Code Injection vulnerability in flash-attn. The records list no fixed version yet; check the vendor advisory below and apply the mitigations in the meantime.

flash-attention contains an insecure deserialization vulnerability in its checkpoint loading mechanism

Also known asPYSEC-2026-2477
Published
Updated
Affected
1 pkg
Patched
None yet
Exploits
None indexed
Exploitation data as of Nov 9, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

Exploitation Status

No confirmed exploitation observed yet

  • CISA’s own triage has not observed active exploitation or public proof-of-concept code for this CVE as of its last assessment.

Exploitation and automatability from CISA’s SSVC triage for CVE-2026-31253.

EPSS Exploitation Probability

via FIRST.org ↗
0.2%probability of exploitation in next 30 days
Lower Risk0.00%
Lower risk than most CVEs12th percentile — riskier than 12% of all scored CVEsHighest risk
0.00%0.29%0.58%0.87%0.0%0.2%Jun 26Sep 26Nov 26

Probability of exploitation in the next 30 days, from FIRST.org EPSS.

How urgent is this, really

Where this sits among everything scored

Real-World Exposure

1 pkg affected
🐍flash-attn

Real-time download stats are indexed for npm and PyPI packages. This vulnerability affects PyPI packages — download data is not available via public APIs for these ecosystems.

Description

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 Packages

1 total
EcosystemPackageVulnerable rangeFix
🐍PyPIflash-attnall versionsNo fix

Detection & mitigation playbook

Open-source dependency

  1. Detect

    Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for flash-attn, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.

  2. Remediation status

    The records we use list no patched version of flash-attn for CVE-2026-31253. Check the maintainer's advisory, and where your build allows, override or pin the dependency away from the vulnerable range.

  3. Mitigate without a patch

    Stop passing untrusted input into the interpreter or shell: call the affected binary with an argument array rather than a composed command string, reject anything outside a strict allowlist of expected values, and run the component under an account that cannot reach beyond the work it legitimately does.

Frequently Asked Questions

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 warms
O3 Security · Impact-Aware SCA

Is CVE-2026-31253 in your dependencies?

Find it across PyPI, including transitive dependencies.

Scan my dependencies How O3 SCA works
CVE-2026-31253: flash-attn RCE (High 7.3)