CVE-2026-22807

Summary

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 Software

VendorProductVersion RangeStatus
vllm-projectvllm>= 0.10.1, < 0.14.0affected

Weaknesses

  • CWE-94: CWE-94: Improper Control of Generation of Code ('Code Injection')

ADP Enrichment

CISA ADP Vulnrichment

  • SSVC:
  • Exploitation: none
    • Automatable: no
    • Technical Impact: total

vLLM: vLLM: Arbitrary code execution via untrusted model loading

Additional References

References