CVE-2026-44223
vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters
描述
vLLM is an inference and serving engine for large language models (LLMs). From to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.
如何修補 CVE-2026-44223
要修補 CVE-2026-44223,請將受影響套件升級到下列已修補版本。
- —升級至 0.20.0 或更新版本
- —升級至 0.20.0 或更新版本
CVE-2026-44223 正在被利用嗎?
低 — EPSS 為 0.4%,目前沒有觀察到大規模利用活動。
受影響套件(2)
- >= 0.18.0, < 0.20.0
- >= 0.18.0, < 0.20.0
CVSS 分數
| 來源 | 版本 | 嚴重程度 | 向量 |
|---|---|---|---|
| osv | CVSS 3.1 | MEDIUM6.5 | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H |