CVE-2026-4372

Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers

Published May 24, 2026 CVSS 3.0 · 7.8 High EPSS < 1% KEV · Not listed
Description
A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker's repository with the victim's full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue.

Remediation

No solution or workaround provided in the CVE record

Tracking

Sign in to view the affected projects.

Advisories

Source ID Title
Github GHSA Github GHSA GHSA-29pf-2h5f-8g72 HuggingFace transformers vulnerable to remote code execution
History
Type Values Removed Values Added
Weaknesses CWE-502
CPEs cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'none', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
First Time appeared Huggingface
Huggingface transformers
Vendors & Products Huggingface
Huggingface transformers
Type Values Removed Values Added
Description A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker's repository with the victim's full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue.
Title Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers
Weaknesses CWE-1066
References
Metrics cvssV3_0

{'score': 7.8, 'vector': 'CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H'}

Weaknesses

Sources

Status
PUBLISHED
Assigner
@huntr_ai
Published
Updated
May 26, 2026, 3:21 PM
Reserved
Mar 18, 2026, 10:04 AM
Status
Analyzed
Published
May 24, 2026, 2:16 PM
Modified
Jul 23, 2026, 5:10 PM