VulnerabilityModified

CVE-2026-41523

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…

HIGH 7.5EPSS 0.91%

Does this matter?

High impact if exploited, but EPSS currently rates exploitation as unlikely (0.91%). Schedule it in the normal patch cycle and watch for a rise in EPSS or a public exploit.

Description

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.

CVSS 3.1: 7.5 HIGHCVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
EPSS: 0.91% probability · 58th percentile
CISA KEV: Not listed
Weakness: CWE-94, CWE-617
Affected: vllm/vllm
Source: security-advisories@github.com

References

Source: NVD record, EPSS from FIRST.org, KEV from CISA. Refreshed daily.