CVE intelligence and bounded remediation
CVE-2026-44223 — vLLM is an inference and serving engine for large language models (LLMs)
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.
- Severity
- Medium
- CVSS
- 6.5 (3.1)
- Published
- 2026-05-12
- CISA KEV
- Not currently listed
- Ecosystem
- software/application
- Weaknesses
- CWE-131, CWE-704
Affected products
- vllm / vllm
Matched remediation archetype
Buffer bounds, memory safety, and memory corruption
This catalog composition supplies bounded fallback guidance. Explicitly reviewed curated workflows load with the complete record below.
Check exposure
- Identify affected native-code versions, build flags, architectures, parsers, codecs, drivers, and input paths in all shipped artifacts.
- Determine whether untrusted data reaches the affected routine and the process privilege, sandbox, and network exposure.
- Confirm statically linked, vendored, firmware, and platform-provided copies, not only package-manager records.
Remediate safely
- Apply the maintained upstream correction or replace the affected component, then rebuild every dependent artifact from clean inputs.
- Adopt bounds-checked interfaces, validated sizes and integer conversions, clear ownership, and memory-safe components where practical.
- Enable supported compiler and runtime hardening and add sanitized tests and fuzz regression seeds derived from non-weaponized fixtures.
Authoritative sources
Complete CVE record and remediation plan
The detailed catalog view below loads this exact record, its source evidence, and the full seven-phase agentic change plan.