Speech-to-text audio decode duration limit bypass via forged header sample rate
CVE-2026-90555 · GHSA-99f2-hwrc-gvq8 · published 2026-08-28 · updated 2026-09-16
- Severity
- Medium
- CVSS score
- 6.5
- Chance of exploitation (EPSS)
- Not scored
- Fixed in
- 0.28.0
- Exploitation No known exploitation
- Automatable No
- Technical impact Partial
- CISA SSVC
Worst realistic outcome: an authenticated client crashes the API-server process (all tenants) with a single ≤3.2 MiB request, repeatably. The decode runs in a ThreadPoolExecutor(max_workers=2) inside the uvicorn process, so two concurrent ceiling-sized requests both saturate the decode pool and commit ≥23.4 GiB RAM (2 × the exact, deterministic 11.7 GiB f.frames × f.channels × 4 primary-buffer figure alone). Preconditions: transcription-capable model served (/v1/audio/* mounted); vllm[audio] installed (soundfile/libsndfile); valid --api-key; no fronting proxy with body cap < ~200 KB. Blast radius: API-server process heap; service-wide DoS. Persistence: none, repeatable on each request.
Upgrade vllm to 0.28.0 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.
| Package | Registry | Affected | Fixed in |
|---|---|---|---|
| vllm | PyPI | < 0.28.0 | 0.28.0 |
- 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 (published by vLLM). Proof-of-concept and reproduction sections and code are left out; the original advisory has them.