CVE-2026-114049

3.7 LOWpublic exploit available

Published 2026-10-21 · Updated 2026-10-22

AI analysis for this CVE has not been generated yet. Raw NVD data is shown below.

NVD description

vLLM before 0.31.0 contains a cross-request mask leak in InputBatch.condense() that allows remote attackers to constrain other clients' outputs by leaving stale allowed_token_ids rows. Attackers can submit concurrent constrained generate requests, including via unauthenticated /invocations, so recycled batch slots apply their token allowlist to victims' unrestricted requests.

CVSS vector

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

Weaknesses

CWE-440

Public exploit & PoC references

https://github.com/vllm-project/vllm/security/advisories/GHSA-6cxc-2vcg-w5qc

All references

https://github.com/vllm-project/vllm/security/advisories/GHSA-6cxc-2vcg-w5qc
https://www.vulncheck.com/advisories/vllm-before-0.31.0-stale-allowed-token-ids-mask-via-inputbatch-condense

Source data: NVD (nvd.nist.gov), public domain. Exploit-DB.ai adds local AI analysis for defensive use only.

Related CVEs

LOWCVE-2026-35191PoC
HIGHCVE-2026-93675
UNSCOREDCVE-2026-97687PoC

Related by shared AI tags and CWE weakness class. Browse the full archive.