CVE record · 1 source
CVE-2026-105752
CVSS 3.1vLLM is an inference and serving engine for large language models.
- Product
- AI / LLM tooling
- CVSS
- 3.1
- Exposure class
- AI infrastructure
Coverage
1Every item in the archive that references this CVE, highest rank first. Headlines open the publisher.
-
vLLM is an inference and serving engine for large language models.
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…