Aller au contenu
en Se connecter
7.8 Élevée

CVE-2026-80047 A vulnerability in Hugging Face Transformers (versions 4.57.0 to 5.16.1) allows remote Python files to be written to…

Description

A vulnerability in Hugging Face Transformers (versions 4.57.0 to 5.16.1) allows remote Python files to be written to local disk without user consent when using GenerativePreTrainedModel.load_custom_generate(). The function fetches and caches a remote module file before performing the required trust_remote_code consent check, inverting the security model enforced by other code-loading paths (such as AutoConfig, AutoModel, and AutoTokenizer). As a result, attacker‑controlled Python code from custom_generate/generate.py is copied into the user’s ~/.cache/huggingface/modules directory even if the user declines the trust prompt. Although execution is correctly gated, the file write is not reversible and can persist across sessions. This can lead to persistent, unauthorized files on disk and stale cache collisions where cached attacker code may later be executed during trusted model loads. The issue stems from an unconditional file write in dynamic_module_utils.py prior to any trust verification.

Produits et versions touchés

Produit Versions touchées Corrigé en
Transformers 4.57 4.57.0 → 5.16.1 Pas de correctif indiqué

Références

Chronologie

  1. Publiée
  2. Dernière modification

Autres CVE de Transformers

Les 12 CVE de Transformers →
CVE-2025-14930 Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability 7.8 Élevée CVE-2025-14928 Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability 7.8 Élevée CVE-2025-14924 Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability 7.8 Élevée CVE-2025-14920 Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability 7.8 Élevée
Suivre cette CVE