CVE-2026-105752
vLLM is an inference and serving engine for large language models.
Does this matter?
Lower severity and a low EPSS score (0.22%). Track it; it rarely justifies an emergency change on its own.
Description
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, Harmony tool continuations submitted through "POST /v1/responses" requests rebuild the next-turn engine input without preserving the cache_salt value, placing the continuation prefix in the global unsalted cache namespace even when the caller enabled salting. On deployments with prefix caching enabled, which is the default, an authenticated tenant who can reconstruct a victim's low-entropy post-tool history can submit the same continuation and use the cached_tokens_per_turn count to determine whether the prefix was previously processed, defeating the intended tenant isolation of salted prefix caching. This issue is fixed in version 0.30.0.
- CVSS 3.1
- 3.1 LOWCVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N
- EPSS
- 0.22% probability · 12th percentile
- Public exploits
- None in Exploit-DB
- CISA KEV
- Not listed
- Weakness
- CWE-200, CWE-524
- Affected
- vllm/vllm
- Source
- security-advisories@github.com
EPSS moved from 0.2% on 6 October 2026 to 0.2% on 26 October 2026.
References
- https://github.com/vllm-project/vllm/commit/6a2a2bb02b563b83f946012959fd3927984d072aPatch
- https://github.com/vllm-project/vllm/pull/50195Issue Tracking, Patch
- https://github.com/vllm-project/vllm/pull/51818Issue Tracking, Patch
- https://github.com/vllm-project/vllm/releases/tag/v0.30.0Release Notes
- https://github.com/vllm-project/vllm/security/advisories/GHSA-935w-9g4m-p28pMitigation, Vendor Advisory
Source: NVD record, EPSS from FIRST.org, KEV from CISA, exploits from Exploit-DB. Refreshed daily. Download this record as JSON.