CVE-2026-31239
Critical 9.8No known exploitationFix status unknownmamba language model framework vulnerable to insecure deserialization when loading pre-trained models from HuggingFace Hub
Vendor advisory: GHSA-pq2f-x424-6fjm
Record last updated
CVSS score
Critical on a 0 to 10 scale
EPSS, chance of exploitation in 30 days
No EPSS score available
CISA remediation deadline
Not in the KEV catalog
Description
The mamba language model framework thru 2.2.6 is vulnerable to insecure deserialization (CWE-502) when loading pre-trained models from HuggingFace Hub. The MambaLMHeadModel.from_pretrained() method uses torch.load() to load the pytorch_model.bin weight file 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 publishing a malicious model repository on HuggingFace Hub. When a victim loads a model from this repository, arbitrary code is executed on the victim's system in the context of the mamba process.
Weakness: CWE-502
Patching and mitigation
- Patch status
- Unknown: no source documents a fix
- Mitigation
- None documented
- Vendor guidance
- github.com (opens in a new tab)
Affected products
| Vendor | Product | Affected versions | Fixed in | Per |
|---|---|---|---|---|
| pip | mamba-ssm | all versions up to and including 2.2.6 | Not documented | GitHub Advisory Database |
Timeline
Disclosure
CVE record published by mitre
Source: CVE Program (opens in a new tab)