CVE-2026-114049 — Vulnerability intelligence
vllm — vllm
Published:
- CVSS
- 6.3
- CISA KEV
- Not listed; not proof of no exploitation
Original source text (not a translation)
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.