8.8

CVE-2026-54232: vllm-project vllm Uncontrolled Search Path Element

CVSS: 8.8v3.1HighAnalyzed
AliasesPYSEC-2026-227GHSA-jrf6-vqxq-pjv2
Published 22-06-2026Last modified 23-06-2026

Vulnerability summary

CVE-2026-54232 is a high (CVSS 8.8) Uncontrolled Search Path Element vulnerability in vllm-project vllm. EPSS estimates a 0.56% chance of exploitation in the next 30 days, higher than 45.9% of scored vulnerabilities. 1 public exploit is recorded.

Overall risk
45/100
0100
Medium · default modelHow is this calculated?
CVSS score
8.8/10
010
High · v3.1 (CVE.org)
EPSS score
0.56%
0%100%
Low · 0.00 pp in 30 days
KEVNot listedIn no KEV catalog
RansomwareNone recordedNo known use recorded
Public exploits1github

Timeline

  1. 22-06-2026
    PublishedVulnerability first disclosed.
  2. 23-06-2026
    Last modifiedVulnerability information updated.

Description

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.1, the vLLM Dockerfile is vulnerable to a dependency confusion attack through the flashinfer-jit-cache package. The package is installed from a custom index (flashinfer.ai/whl/) using --extra-index-url, but the package name was not registered on PyPI, and UV_INDEX_STRATEGY="unsafe-best-match" is set globally. An attacker who registers flashinfer-jit-cache on PyPI with version 0.6.11.post2 can execute arbitrary code as root during the Docker build and backdoor every resulting container image, enabling exfiltration of all user prompts, API credentials, and model data from production vLLM deployments This vulnerability is fixed in 0.22.1.

CVSS metrics

3 scores · 3 sources · 1 version

EPSS trend

Probability

0.56%

Chance of exploitation in the next 30 days

Percentile

46%

Versus all other CVEs

30-day change

0.00 pp

From 0.56% on 21-09-2026

21-09-202621-10-2026
EPSS score
0.56%

Known exploited

Compare KEV catalogs

Not in any of the 3 tracked catalogs.

Public exploits

1 found
  • githubUnverified

    Vllm

    Maturity: poc

Weaknesses

CWE · root cause

Affected systems

PyPI1 product
  • vllm< 0.22.1
Vllm1 product
Vllm-project1 product

3 vendors · 3 products

References

1 links