vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing a malicious HuggingFace model, when vLLM runs in Python optimized mode (python -O or PYTHONOPTIMIZE=1). This vulnerability is fixed in 0.22.0.
01
describes (1)
Related CVEs.
| ID | severity | exploitable | ||
|---|---|---|---|---|
| CVE-2026-41523US | — | — | ||
vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthent… | ||||
02
references
External citations — patches, advisories, analysis.
- https://access.redhat.com/errata/RHSA-2026:36005ADVISORY
- https://access.redhat.com/errata/RHSA-2026:36006ADVISORY
- https://access.redhat.com/security/cve/CVE-2026-41523WEB
- https://bugzilla.redhat.com/show_bug.cgi?id=2491582REPORT
- https://github.com/vllm-project/vllm/commit/b3c7ffcab82c2439726f8cb213800f6f38c023d3FIX
- https://github.com/vllm-project/vllm/security/advisories/GHSA-q8gq-377p-jq3rEVIDENCE
- https://huntr.com/bounties/dcb05b04-e625-41e7-adbc-bbae0cc2d64cADVISORY
- https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-41523.jsonWEB
scry
Advisories like this name the exposed products attackers go after on your perimeter. See what is exposed on your perimeter with a free passive scan.
Free scan