---
id: CVE-2026-105755
title: vLLM is an inference and serving engine for large language models
summary: >-
  vLLM is an inference and serving engine for large language models. Prior to
  0.30.0, flash late-interaction scoring at the /score and /rerank endpoints
  derives each worker's query_key value from the caller-controlled X-Request-Id
  header. …
severity: medium
cvss: 4.2
cvssVector: 'CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L'
cwe:
  - CWE-639
  - CWE-694
vendor: vllm
product: vllm
affected:
  - vllm < 0.30.0
patched:
  - vllm 0.30.0
published: '2026-10-05'
updated: '2026-10-08'
sourceUpdated: '2026-10-08T01:28:50.393'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-105755'
references:
  - url: >-
      https://github.com/vllm-project/vllm/commit/ee17d0d869203ef9a35ad73358a4987bba14b1fc
    label: security-advisories@github.com
  - url: 'https://github.com/vllm-project/vllm/pull/51445'
    label: security-advisories@github.com
  - url: 'https://github.com/vllm-project/vllm/releases/tag/v0.30.0'
    label: security-advisories@github.com
  - url: >-
      https://github.com/vllm-project/vllm/security/advisories/GHSA-2phq-3phc-84px
    label: security-advisories@github.com
  - url: 'https://github.com/advisories/GHSA-2phq-3phc-84px'
  - url: 'https://github.com/vllm-project/vllm'
  - url: >-
      https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-105755.json
  - url: 'https://access.redhat.com/security/cve/CVE-2026-105755'
  - url: 'https://bugzilla.redhat.com/show_bug.cgi?id=2546184'
  - url: 'https://www.cve.org/CVERecord?id=CVE-2026-105755'
  - url: 'https://nvd.nist.gov/vuln/detail/CVE-2026-105755'
tags:
  - nvd
  - cve.org
  - ghsa
  - pip
  - osv
  - csaf
  - vex
  - red-hat
epss: 0.00202
epssPercentile: 0.10094
aliases:
  - GHSA-2phq-3phc-84px
  - PYSEC-2026-4198
ecosystem: pip
ssvc:
  exploitation: none
  automatable: 'no'
  technicalImpact: partial
  timestamp: '2026-10-06T17:55:11.724369Z'
ingestedAt: '2026-10-05T23:36:21.156Z'
---

## Overview

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0.

## Affected

- `vllm < 0.30.0`

## Remediation

Upgrade past the affected range:

- `vllm 0.30.0`

## Package advisory (CVE-2026-105755)

Affected packages:

- `vllm < 0.30.0`

Patched in:

- `vllm 0.30.0`

Source: https://github.com/advisories/GHSA-2phq-3phc-84px

## Vendor advisories

- **Red Hat VEX** · Moderate · affected: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · no fix planned: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · updated 2026-10-06 · [vex](https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-105755.json)