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.
Published: 2026-05-12
Score: 9.8 Critical
EPSS: < 1% Very Low
KEV: No
Impact: n/a
Action: n/a
AI Analysis

Analysis and contextual insights are available on OpenCVE Cloud.

Remediation

No vendor fix or workaround currently provided.

Additional remediation guidance may be available on OpenCVE Cloud.

Tracking

Sign in to view the affected projects.

Advisories

No advisories yet.

History

Thu, 14 May 2026 23:00:00 +0000

Type Values Removed Values Added
Title Insecure Deserialization in Mamba Language Model Framework 2.2.6 Allows Remote Code Execution

Thu, 14 May 2026 20:30:00 +0000

Type Values Removed Values Added
Metrics cvssV3_1

{'score': 9.8, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H'}


Wed, 13 May 2026 11:00:00 +0000

Type Values Removed Values Added
First Time appeared State-spaces
State-spaces mamba
Vendors & Products State-spaces
State-spaces mamba

Tue, 12 May 2026 18:45:00 +0000

Type Values Removed Values Added
Title Insecure Deserialization in Mamba Language Model Framework 2.2.6 Allows Remote Code Execution
Weaknesses CWE-502

Tue, 12 May 2026 17:30:00 +0000

Type Values Removed Values Added
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.
References

Subscriptions

State-spaces Mamba
cve-icon MITRE

Status: PUBLISHED

Assigner: mitre

Published:

Updated: 2026-05-14T19:54:17.202Z

Reserved: 2026-03-09T00:00:00.000Z

Link: CVE-2026-31239

cve-icon Vulnrichment

No data.

cve-icon NVD

Status : Deferred

Published: 2026-05-12T18:16:52.320

Modified: 2026-05-14T20:17:03.593

Link: CVE-2026-31239

cve-icon Redhat

No data.

cve-icon OpenCVE Enrichment

Updated: 2026-05-14T22:45:31Z

Weaknesses