CVE-2026-105752
vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle
Auswirkung
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.
Ein Angreifer kann die Lücke über das Netzwerk ausnutzen, mit einem normalen Benutzerkonto, ohne Zutun eines Nutzers. Dafür müssen bestimmte Bedingungen erfüllt sein, was einen Angriff erschwert. Bei Erfolg: geringes Risiko, dass Daten oder Systeme verändert werden.
Betroffene Pakete
| Ökosystem | Paket | Betroffen | Behoben in |
|---|---|---|---|
| pip | vllm | < 0.30.0 | 0.30.0 |
Betroffene Produkte · Vllm
| Produkt | Betroffene Versionen |
|---|---|
| Vllm | vor 0.30.0 |
Schwachstellenklasse
Referenzen
- NVD-Eintrag
- CVE.org-Eintrag
- GitHub-Advisory
- https://github.com/advisories/GHSA-935w-9g4m-p28p
- https://github.com/vllm-project/vllm/security/advisories/GHSA-935w-9g4m-p28p
- https://github.com/vllm-project/vllm/pull/50195
- https://github.com/vllm-project/vllm/pull/51818
- https://github.com/vllm-project/vllm/commit/6a2a2bb02b563b83f946012959fd3927984d072a
- https://github.com/vllm-project/vllm/releases/tag/v0.30.0