All vulnerabilities

Servers and applications

CVE-2026-44223

extract_hidden_states speculative decoding crashes server on any request with penalty parameters

Severity
Medium 6.5
Published
28 Apr 2026
Last changed
11 Nov 2026
Added to UpHeron
3 Nov 2026
Fixed in
0.20.0
Exploit chance
0.4% in the next 30 daysHigher than 31% of CVEs. EPSS by FIRST, 19 Nov 2026
Weakness
CWE-131Incorrect type conversion (CWE-704)
Attack
Over the network · Needs a user account · No user action needed
Attacker can
Take it down
CVSS vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
Also known as
GHSA-83vm-p52w-f9pw
Sources
NVD, Project advisories (GitHub)

Description

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. The crash is deterministic and immediate — no concurrency, race condition, or special workload is required.

In vLLM v0.17.0, the extract_hidden_states proposer's propose() method returned sampled_token_ids.unsqueeze(-1), producing a tensor of shape (batch_size, 1).

In PR #37013 (first released in v0.18.0), the KV connector interface was refactored out of propose(). The return type changed from tuple[Tensor, KVConnectorOutput | None] to Tensor, and the .unsqueeze(-1) call was removed along with the KV connector output:

From vllm-project/vllm's security advisory. Shortened; the full text is there.

Workarounds

  • Upgrade to vLLM v0.20.0 or later.
  • If upgrading is not possible, avoid using extract_hidden_states as the speculative decoding method on affected versions.
  • Alternatively, reject or strip penalty parameters (repetition_penalty, frequency_penalty, presence_penalty) from incoming requests at an API gateway before they reach vLLM.
Workarounds as given by vllm-project/vllm's security advisory.

Affected software

SoftwareAffected versions
vLLMServers and applications · cpe:a:vllm:vllm
  • From 0.18.0 before 0.20.0
vLLMServers and applications · github:vllm-project/vllm
  • From 0.18.0 before 0.20.0

About vLLM

open-source software for large language model inference

From Wikidata, CC0 1.0.
Docker Hub pulls
44,195,225Docker Hub, 15 Nov 2026
GitHub stars
96,943GitHub, 15 Nov 2026
Forks
24,836GitHub, 15 Nov 2026
Latest release
v0.33.0rc3GitHub, 15 Nov 2026
Released
2 Nov 2026GitHub
License
Apache Software License 2.0Wikidata
First published on GitHub
9 Feb 2023GitHub

References

  • NVD https://nvd.nist.gov/vuln/detail/CVE-2026-44223
  • CVE.org https://www.cve.org/CVERecord?id=CVE-2026-44223
  • GitHub Advisory Database https://github.com/advisories/GHSA-83vm-p52w-f9pw
  • github.com https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw