CVE-2026-22807: vLLM affected by RCE via auto_map dynamic module loading during model initialization

Severity: critical (CVSS 9.8 v3.1)
Exploited in the wild: not listed in CISA KEV
EPSS: 1.0% probability of exploitation in the next 30 days
Public exploit or detection template: none known
Published: 2026-01-21 · updated 2026-07-21
Weakness: CWE-94

Description

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.

Affected products

vllm-project vllm: from 0.10.1 before 0.14.0 — fixed in 0.14.0

How to fix

Upgrade to a fixed version: vllm-project vllm 0.14.0. Check the vendor advisory for mitigations.

References

Get alerted when CVE-2026-22807 or the next vulnerability affects your stack