Use of Incorrectly-Resolved Name or Reference Affecting vllm package, versions [,0.22.0)
Threat Intelligence
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
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-PYTHON-VLLM-17304846
- published11 Jun 2026
- disclosed10 Jun 2026
- creditaddcontent
Introduced: 10 Jun 2026
NewCVE-2026-47155 (opens in a new tab)How to fix?
Upgrade vllm to version 0.22.0 or higher.
Overview
vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs
Affected versions of this package are vulnerable to Use of Incorrectly-Resolved Name or Reference through several model loading paths. An attacker can make the server load a different Hugging Face artifact than intended by supplying a model reference that is fetched without the requested revision, or by triggering secondary artifact loads that omit the pinned revision or code revision. This can cause the user to run or serve the wrong model weights and associated remote code, leading to model integrity loss and exposure to attacker-controlled model behavior.