Instituto Nacional de ciberseguridad. Sección Incibe
Instituto Nacional de Ciberseguridad. Sección INCIBE-CERT

CVE-2026-22807

Gravedad CVSS v3.1:
ALTA
Tipo:
CWE-94 Control incorrecto de generación de código (Inyección de código)
Fecha de publicación:
21/01/2026
Última modificación:
21/01/2026

Descripción

*** Pendiente de traducción *** vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.14.0, vLLM loads Hugging Face `auto_map` dynamic modules during model resolution without gating on `trust_remote_code`, allowing attacker-controlled Python code in a model repo/path to execute at server startup. An attacker who can influence the model repo/path (local directory or remote Hugging Face repo) can achieve arbitrary code execution on the vLLM host during model load. This happens before any request handling and does not require API access. Version 0.14.0 fixes the issue.