GCVE Workshop - 22 September 2026 (14:00-18:00), Luxembourg Before The Vulnopticon Conference - Registration

BREW-MLX-LM-CVE-2026-5241 (GHSA-FGCW-684Q-JJ6R)

Vulnerability from osv_homebrew – Published: 2026-08-13 17:14 – Updated: 2026-09-10 00:32 – Source website
VLAI
Summary
huggingface/transformers: Arbitrary Code Execution During Model Initialization in the LightGlue Model Loading Path
Details

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.

Severity
8.0 (High)
CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N
References
To clipboard 

{
  "affected": [
    {
      "ecosystem_specific": {
        "fix": "bump",
        "range_state": "fixed",
        "resource": "transformers",
        "resource_purl": "pkg:pypi/transformers@5.16.1",
        "upstream_fixed_in": "5.5.0"
      },
      "package": {
        "ecosystem": "Homebrew",
        "name": "mlx-lm",
        "purl": "pkg:brew/mlx-lm"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.31.3_2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "database_specific": {
    "confidence": "high",
    "source": "matched",
    "strategy": "registry",
    "upstream_evidence": [
      {
        "ecosystem": "PyPI",
        "key": "pkg:pypi/transformers@5.16.1",
        "name": "transformers",
        "resource": "transformers",
        "strategy": "registry",
        "subject_version": "5.16.1"
      }
    ]
  },
  "details": "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.",
  "id": "BREW-mlx-lm-CVE-2026-5241",
  "modified": "2026-09-10T00:32:14Z",
  "published": "2026-08-13T17:14:17Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-5241"
    },
    {
      "type": "WEB",
      "url": "https://github.com/huggingface/transformers/commit/676559d5022b74aaa0cee1cee0842b7f27c5320e"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:34456"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:37275"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42644"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/security/cve/CVE-2026-5241"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2484384"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/huggingface/transformers"
    },
    {
      "type": "WEB",
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/transformers/PYSEC-2026-2290.yaml"
    },
    {
      "type": "WEB",
      "url": "https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619"
    },
    {
      "type": "WEB",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-5241.json"
    }
  ],
  "schema_version": "1.7.3",
  "severity": [
    {
      "score": "CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "huggingface/transformers: Arbitrary Code Execution During Model Initialization in the LightGlue Model Loading Path",
  "upstream": [
    "GHSA-fgcw-684q-jj6r",
    "CVE-2026-5241",
    "PYSEC-2026-2290"
  ]
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Detection rules are retrieved from Rulezet.


Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.