LMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.py
CVE-2025-66455 · GHSA-2vh9-42vm-xmv2 · published 2026-09-18 · updated 2026-09-21
- Severity
- Critical
- CVSS score
- 9.8
- Chance of exploitation (EPSS)
- 1.3%
- Fixed in
- 0.16.0
- Exploitation No known exploitation
- Automatable Yes
- Technical impact Total
- CISA SSVC
Successful exploitation allows arbitrary code execution as the LMDeploy serving process. This can expose model weights, prompts, credentials, attached storage, cluster-network services, and host or GPU resources. An attacker may also modify or terminate the serving process.
Upgrade lmdeploy to 0.16.0 or later.
The issue was fixed by replacing the pickle-based ZeroMQ protocol with JSON serialization: send_pyobj() was replaced with send_json(). recv_pyobj() was replaced with recv_json(). Received objects are validated using the DistServeCacheFreeRequest Pydantic schema before use. Invalid or off-schema messages are rejected without terminating the receive loop. Fix commit: https://github.com/InternLM/lmdeploy/commit/f05b4ad8bf2e2d84101a1d63b3c44fadd99223b2 The fix was released in LMDeploy 0.16.0.
Users who cannot upgrade immediately should: Prevent untrusted clients from reaching /distserve/* endpoints. Restrict the DistServe HTTP and ZeroMQ control planes to trusted cluster networks. Configure API-key authentication. Block arbitrary outbound ZeroMQ connections from serving nodes. These measures reduce exposure but do not make pickle deserialization safe. Upgrading to LMDeploy 0.16.0 or later is recommended.
| Package | Registry | Affected | Fixed in |
|---|---|---|---|
| lmdeploy | PyPI | >= 0.9.2, < 0.16.0 | 0.16.0 |
- 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
- Criticaln8n has XML Node Prototype Pollution that to RCEn8n, n8n, n8n
- Criticaln8n has Prototype Pollution in XML Webhook Body Parser that Leads to RCEn8n, n8n, n8n
- CriticalLMdeploy has Remote Code Execution by Pickle Deserialization via zmq_rpc.call_and_response() in InterLM/lmdeploylmdeploy
- HighLMDeploy has an SSRF bypasslmdeploy
- HighLMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loadinglmdeploy
- Highlmdeploy: Hardcoded trust_remote_code=True is an implicit unsafe remote-code load path with no user opt-outlmdeploy
- HighLMDeploy: Arbitrary code execution via hardcoded trust_remote_code=True in lmdeploy model initializationlmdeploy
- HighLMDeploy has Server-Side Request Forgery (SSRF) via Vision-Language Image Loadinglmdeploy
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.