Arbitrary Code Injection Affecting tritonserver-backend-vllm-cuda-13.0 package, versions <25.11-r2


Severity

Recommended
0.0
critical
0
10

Snyk's Security Team recommends NVD's CVSS assessment. Learn more

Threat Intelligence

Social Trends
EPSS
0.74% (52nd percentile)

Do your applications use this vulnerable package?

In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.

Test your applications
  • Snyk IDSNYK-CHAINGUARDLATEST-TRITONSERVERBACKENDVLLMCUDA130-15855793
  • published31 Mar 2026
  • disclosed21 Jan 2026

Introduced: 21 Jan 2026

CVE-2026-22807  (opens in a new tab)
CWE-94  (opens in a new tab)

How to fix?

Upgrade Chainguard tritonserver-backend-vllm-cuda-13.0 to version 25.11-r2 or higher.

NVD Description

Note: Versions mentioned in the description apply only to the upstream tritonserver-backend-vllm-cuda-13.0 package and not the tritonserver-backend-vllm-cuda-13.0 package as distributed by Chainguard. See How to fix? for Chainguard relevant fixed versions and status.

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.

CVSS Base Scores

version 3.1