CVE-2026-105752 vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle
Authenticated tenants can infer victim activity via prefix cache timing, bypassing tenant isolation in vLLM deployments with default prefix caching enabled.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- Available
What is CVE-2026-105752?
CVE-2026-105752 is a low-severity vulnerability in Vllm Vllm, published on 5 Oct 2026. Any signed-in user can exploit it over the network, with no action from the victim.
Official 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.
Which versions are affected by CVE-2026-105752?
3 open source packages ship an affected version, including vllm. It affects 1 product.
| Package | Affected | Fixed in |
|---|---|---|
vllm PyPI | All versions | 0.30.0 |
vllm PyPI | All versions | 0.30.0 |
vllm PyPI | All versions | 0.30.0 |
Products include Vllm Vllm.
How do I fix CVE-2026-105752?
Upgrade vllm to 0.30.0 or later.
Is CVE-2026-105752 being exploited?
There is no known exploitation of CVE-2026-105752 in the wild and no public exploit code.
Other vulnerabilities in Vllm Vllm
- CVE-2025-32444 vLLM Vulnerable to Remote Code Execution via Mooncake Integration
- CVE-2026-22778 vLLM leaks a heap address when PIL throws an error
- CVE-2025-47277 vLLM Allows Remote Code Execution via PyNcclPipe Communication Service
- CVE-2024-11041 Remote Code Execution in vllm-project/vllm
- CVE-2026-48746 vLLM: OpenAI auth bypass
Sources
Also known as GHSA-935w-9g4m-p28p, MINI-cfq3-w5c9-chgv, MINI-j935-7xrw-x32q, MINI-mf8m-h427-m656, PYSEC-2026-4308.