CVE-2026-5241: transformers — Exploit & Fix

Description

A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment.

Is your site exposed to CVE-2026-5241?

Run a free security scan — no signup, results in seconds.

CVSS v3.1 Score

9.6

CRITICAL

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H

EPSS — Exploit Prediction

0.0049

Probability of exploitation

0.38%

Percentile rank

EPSS estimates the probability that this vulnerability will be exploited in the wild within the next 30 days. A higher score means more likely to be exploited.

Weakness Type (CWE)

CWE-829 CWE-829

Affected Products

Vendor	Product	Versions
huggingface	transformers	All

References

Advisories & Patches

https://github.com/huggingface/transformers/commit/676559d5022b74aaa0cee1cee0842b7f27c5320e

Exploits

https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619 https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619

Frequently Asked Questions

What is CVE-2026-5241? +

A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment. It has a CVSS v3.1 base score of 9.6 (CRITICAL).

How severe is CVE-2026-5241? +

CVE-2026-5241 has a CVSS v3.1 score of 9.6 out of 10, rated CRITICAL. This is a critical vulnerability that should be patched immediately. The EPSS score is 0.0049, placing it in the 0th percentile for exploitation probability.

What products are affected by CVE-2026-5241? +

CVE-2026-5241 affects products from huggingface, specifically: transformers. Check the affected products table above for specific version ranges.

How do I check if I'm vulnerable to CVE-2026-5241? +

You can use Secably's free Website Scanner to check your website for known vulnerabilities. For infrastructure scanning, use the Port Scanner to identify exposed services that may be affected. Check the vendor advisories linked above for specific patch and version information.

Related Vulnerabilities

CVE-2025-34060

A PHP objection injection vulnerability exists in the Monero Project’s Laravel-based forum software due to unsafe handling of untrusted input …

CVE-2026-96269

GNU Emacs 28.1 through 31.1 allows arbitrary code execution upon opening a file, because an untrusted value of read-symbol-shorthands affects …

CVE-2025-34074

An authenticated remote code execution vulnerability exists in Lucee’s administrative interface due to insecure design in the scheduled task functionality. …

CVE-2025-36852

A critical security vulnerability exists in remote cache extensions for common build systems utilizing bucket-based remote cache (such as those …

CVE-2026-6357

pip prior to version 26.1 would run self-update check functionality after installing wheel files which required importing well-known Python modules …

CVE-2023-7357

A remote code execution (RCE) vulnerability was identified in the FlexNet Publisher lmadmin web user interface. This vulnerability is addressed …