Dependency Confusion Vulnerability in vLLM by vLLM Project
CVE-2026-54232

8.8HIGH

Key Information:

Status
Vendor
CVE Published:
22 June 2026

What is CVE-2026-54232?

The vLLM inference and serving engine for large language models has a vulnerability related to dependency confusion arising from the installation of the 'flashinfer-jit-cache' package from a non-PyPI source. This misconfiguration exposes vLLM deployments to significant risks, as an attacker can exploit this by publishing a malicious version of the package on PyPI, enabling them to execute arbitrary code with root privileges during Docker image builds. Consequently, this not only compromises the integrity of the container images generated but also risks the exfiltration of sensitive data including user prompts and API credentials. The vulnerability is addressed in the release of vLLM version 0.22.1.

Affected Version(s)

vllm < 0.22.1

References

CVSS V3.1

Score:
8.8
Severity:
HIGH
Confidentiality:
High
Integrity:
High
Availability:
High
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
None
User Interaction:
Required
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.