Vulnerability Description
PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which is commonly used to load saved model states, internally calls torch.load() without setting the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, leading to arbitrary code execution on the victim's system when the file is loaded.
CVSS Score
HIGH
Affected Products
| Vendor | Product | Versions |
|---|---|---|
| Lightningai | Pytorch Lightning | <= 2.6.0 |
Related Weaknesses (CWE)
References
- https://github.com/Lightning-AI/pytorch-lightningProduct
- https://www.notion.so/CVE-2026-31221-35d1e1393188815f8db7c4fd08076639ExploitThird Party Advisory
FAQ
What is CVE-2026-31221?
CVE-2026-31221 is a vulnerability with a CVSS score of 7.8 (HIGH). PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which i...
How severe is CVE-2026-31221?
CVE-2026-31221 has been rated HIGH with a CVSS base score of 7.8/10. Review the CVSS metrics above for detailed severity breakdown.
Is there a patch for CVE-2026-31221?
Check the references section above for vendor advisories and patch information. Affected products include: Lightningai Pytorch Lightning.