CVEFinder.io

CVE-2026-31239

⛔ critical
🔍 Scan for this CVE
Summary

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 H

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.

CVSS Score
9.8
Critical
EPSS Score
0.0
Exploit Probability
Published Date
2026-05-12
First Seen: 2026-05-17
📊 Relative Risk Intelligence

This CVE is Very High Risk - more severe than 90.5% of all 328,009 vulnerabilities in our database.

#31,168
Top 10% most severe
Severity Percentile
🎯 CISA SSVC Assessment Updated: May 14, 2026
🔍 Exploitation Status
None
No known exploits
⚙️ Automatable
YES
Can be exploited automatically
💥 Technical Impact
Total
Complete system compromise possible
SSVC data provided by CISA
Last Modified 2026-05-14
Source NVD 🔗
CVSS Vector 3.1 CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
CWE IDs (Weakness Types)

📦 Affected Products 0

No affected products information available

🔗 References 2