CVE-2026-44223
is CVE-2026-44223real, exploitable, or a false positive? Here's the community verdict.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
vLLM is an inference and serving engine for large language models (LLMs). From 0.18.0 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.
References
Published
Embed this verdict
[](https://www.truepositive.app/cve/CVE-2026-44223)<a href="https://www.truepositive.app/cve/CVE-2026-44223"><img src="https://www.truepositive.app/cve/CVE-2026-44223/badge.svg" alt="TruePositive verdict for CVE-2026-44223"></a>Live badge that updates automatically as the community verdict changes.
Community ground truth
In your experience, is this finding real and exploitable?
0 verdictsNo account needed. Anonymous verdicts post as an unverified signal. Log in to make yours verified and earn reputation.
Field notes & remediation
Verdicts are the quick signal. Notes are the evidence and fixes behind them.
No notes yet. Be the first to share what you saw, or a fix that worked.