CVE-2026-54232

Vllm remote code execution

8.8 HIGHPublic exploitEPSS 0.56%

Published22 Jun 2026

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 3.1 breakdown

Attack vectorNetwork

Attack complexityLow

Privileges requiredNone

User interactionRequired

ScopeUnchanged

ConfidentialityHigh

IntegrityHigh

AvailabilityHigh

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

Affected products

Vllm< 0.22.1

Remediation

Fixed in

Vllm≥ 0.22.1

References

https://github.com/vllm-project/vllm/security/advisories/GHSA-jrf6-vqxq-pjv2ExploitThird Party Advisory

Related CVEs

Recent high-severity vulnerabilities sharing a weakness type.

Frequently asked

Is CVE-2026-54232 being actively exploited?

CVE-2026-54232 is not in the CISA KEV catalog. A public exploit reference exists, so treat it as exploitable. Its EPSS exploit probability is 0.56%.

How severe is CVE-2026-54232?

CVE-2026-54232 has a CVSS 3.1 base score of 8.8, rated high. EPSS estimates a 0.56% probability of exploitation.

Which products are affected by CVE-2026-54232?

Affected products include Vllm.

How do I fix CVE-2026-54232?

Upgrade affected software to a fixed version, for example Vllm 0.22.1 or later, then confirm no affected versions remain in your inventory.

How Secuno handles this

Secuno correlates vulnerabilities like CVE-2026-54232 against the real software inventory on every managed device, every day. Rather than a static severity, each match is scored in context: exploit availability, EPSS, the asset's exposure, and business criticality decide where it lands in the queue. If an affected version is present anywhere across your portfolio, it surfaces automatically with a prioritised remediation path.

See how prioritisation works Book a demo

Data from the NVD, CISA KEV catalog and FIRST EPSS. CVSS 3.1 base score shown where available.