Home

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.

PUBLISHED Reserved 2026-03-31 | Published 2026-06-03 | Updated 2026-06-03 | Assigner @huntr_ai




HIGH: 8.0CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N

Problem types

CWE-829 Inclusion of Functionality from Untrusted Control Sphere

Product status

Any version before 5.5.0
affected

References

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

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

github.com/...ommit/676559d5022b74aaa0cee1cee0842b7f27c5320e

cve.org (CVE-2026-5241)

nvd.nist.gov (CVE-2026-5241)

Download JSON