vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters
CVE-2026-44223 · GHSA-83vm-p52w-f9pw · published 2026-05-06 · updated 2026-06-08
- Severity
- Medium
- CVSS score
- 6.5
- Chance of exploitation (EPSS)
- 0.37%
- Fixed in
- 0.20.0
- Exploitation Public proof of concept
- Automatable No
- Technical impact Partial
- CISA SSVC
Any vLLM deployment between v0.18.0 and v0.19.1 (inclusive) configured with extracthiddenstates speculative decoding is affected. A single API request containing any penalty parameter immediately and permanently crashes the EngineCore process, resulting in complete loss of service availability.
Upgrade vllm to 0.20.0 or later. If you start it from an MCP config with npx or uvx, pin that version so an old cached copy is not used.
Fixed in PR #38610, first included in vLLM v0.20.0. The fix slices the return value to sampledtokenids[:, :1], ensuring the correct (batch_size, 1) shape regardless of the rejection sampler's output dimensions.
Upgrade to vLLM v0.20.0 or later. If upgrading is not possible, avoid using extracthiddenstates as the speculative decoding method on affected versions. Alternatively, reject or strip penalty parameters (repetitionpenalty, frequencypenalty, presence_penalty) from incoming requests at an API gateway before they reach vLLM.
| Package | Registry | Affected | Fixed in |
|---|---|---|---|
| vllm | PyPI | >= 0.18.0, < 0.20.0 | 0.20.0 |
- CriticalLiteLLM has SQL Injection in Proxy API key verificationlitellmExploited
- Criticaln8n Vulnerable to Remote Code Execution via Expression Injectionn8n, n8nExploited
- HighLiteLLM has a sandbox escape in custom-code guardraillitellmExploited
- HighLangflow affected by Remote Code Execution via validate_code() exec()langflowExploited
- CriticalJupyter MCP Server: Browser DNS rebinding can hijack the Jupyter backend through management routesjupyter-mcp-server
- Criticalmem0ai Databricks vector store vulnerable to SQL injectionmem0ai
- CriticalvLLM: OpenAI auth bypassvllm
- CriticalvLLM has RCE In Video Processingvllm
- CriticalvLLM Allows Remote Code Execution via PyNcclPipe Communication Servicevllm
- CriticalvLLM Vulnerable to Remote Code Execution via Mooncake Integrationvllm
- CriticalCVE-2025-24357 Malicious model remote code execution fix bypass with PyTorch < 2.6.0vllm
- CriticalvLLM Deserialization of Untrusted Data vulnerabilityvllm
The impact, patch and workaround text comes from the advisory, shortened (GitHub Advisory Database, CC BY 4.0). Proof-of-concept and reproduction sections and code are left out; the original advisory has them.