CVE-2026-105752
lowvLLM is an inference and serving engine for large language models.
Vendor / product: Vllm — vllm
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/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N Affected products
Weakness classification (CWE)
Source: NVD / CISA. “NVD-CWE-Other” and “NVD-CWE-noinfo” mean NVD could not assign a specific weakness.
Mitigation & advisories
References
- https://github.com/vllm-project/vllm/commit/6a2a2bb02b563b83f946012959fd3927984d072a security-advisories@github.com
- https://github.com/vllm-project/vllm/pull/50195 security-advisories@github.com
- https://github.com/vllm-project/vllm/pull/51818 security-advisories@github.com
- https://github.com/vllm-project/vllm/releases/tag/v0.30.0 security-advisories@github.com
- https://github.com/vllm-project/vllm/security/advisories/GHSA-935w-9g4m-p28p security-advisories@github.com