vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled
CVE-2026-115912 · GHSA-6757-hggm-hp2w · published 2026-10-28
- Severity
- High
- CVSS score
- 7.5
- Chance of exploitation (EPSS)
- 0.51%
- Fixed in
- 0.32.0
- Exploitation No known exploitation
- Automatable No
- Technical impact Total
- CISA SSVC
An ordinary authenticated inference caller can execute arbitrary Python in the worker process on a deployment that uses SHM multimodal caching. The live proof used a benign marker gadget and also caused the EngineCore health endpoint to fall from 200 to 503 after the forged request. The exploit is reachable with VLLMALLOWINSECURE_SERIALIZATION=0; it abuses the built-in SHM object deserializer, not the separate optional msgpack ext-hook flag.
Upgrade vllm to 0.32.0 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.
Authenticate every SHM carrier with a server-issued capability bound to the cache key, address, and monotonic id; rotate the capability secret when the cache is cleared; and reject malformed or forged SHM carriers at the public scale-out boundary before engine handoff — so a forged pair never reaches the worker deref / pickle.loads(). The sender emits an HMAC capability (keyed by a 32-byte secret stored in the ring buffer and rotated on clear()) over (cachekey, address, monotonicid) alongside the pair. The public scale-out path validates it before engine handoff:
| Package | Registry | Affected | Fixed in |
|---|---|---|---|
| vllm | PyPI | < 0.32.0 | 0.32.0 |
- CriticalLMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.pylmdeploy
- CriticalLMdeploy has Remote Code Execution by Pickle Deserialization via zmq_rpc.call_and_response() in InterLM/lmdeploylmdeploy
- CriticalSGLang: Unauthenticated RCE via --enable-custom-logit-processorsglang
- CriticalSGLanG: Multimodal scheduler deserializes untrusted pickle data on 0.0.0.0 ROUTER socketsglang
- Criticaln8n: HTTP Request Node Pagination Prototype Pollution to RCEn8n, n8n, n8n
- Criticaln8n Has an XML Node Prototype Pollution Patch Bypassn8n, n8n, n8n
- 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.