vLLM: Mirrored multimodal IPC caches desync after a rejected request — a later request reusing the same media hash trips a receiver assertio
CVE-2026-105753 · GHSA-ph3r-5jfg-f84f · published 2026-10-06
- Severity
- Medium
- CVSS score
- 6.5
- Chance of exploitation (EPSS)
- 0.43%
- Fixed in
- 0.28.0
- Exploitation No known exploitation
- Automatable No
- Technical impact Partial
- CISA SSVC
A remote client submitting multimodal requests can poison a cache identity — render a media item successfully, then have that request rejected — so a later request reusing the same media hash fails on the P1 receiver assertion. This is an availability failure against a shared serving instance. No code execution, memory corruption, or data disclosure is claimed. On this revision the failure is scoped as a request-level preprocessing error (the engine core catches around preprocessaddrequest); public reports show the same assertion cascading into further engine-loop assertions on other revisions. It applies to multimodal models running the default mirrored lru cache.
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.
Two complementary changes: 1. Make the mirrored commit atomic with admission — insert into the P0 sender cache only after the request has passed all admission checks (length, limits) and P1 has acknowledged the item, or roll back the P0 insert on rejection. 2. Defense in depth — convert the P1 receiver assert mmitem is not None into a checked, request-scoped error (fetch-on-miss from P0) so a desync degrades a single request rather than asserting in the engine loop. The core of the rollback half: wrap the post-render length check so a ValueError rejection discards the P0 entries the render just committed, before re-raising.
| 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 (GitHub Advisory Database, CC BY 4.0). Proof-of-concept and reproduction sections and code are left out; the original advisory has them.