PrivacyHigh
CVE-2026-105755: vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-inter…
- Occurred
- Oct 5, 2026
- Published
- Oct 5, 2026
- Source
- NVD CVE
Summary
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can overwrite the cached query embedding so the victim's documents are scored against the attacker's query, and shared use counters can also cause a late-interaction cache-miss error. This issue is fixed in version 0.30.0.
Tags
- CVE
- CVE-2026-105755
- cvss:MEDIUM
Classification
Category confidence: 90%
Severity confidence: 85%
Rationale: CVE describes a security vulnerability allowing request ID reuse to overwrite cached embeddings, potentially exposing sensitive data or causing unauthorized access.
Classifier: Qwen/Qwen3.5-35B-A3B-FP8