Lmdeploy

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions 012.1 through 0.12.2 contain a code injection vulnerability in `lmdeploy/pytorch/config.py` line 620 that allows an attacker to execute arbitrary Python code by publishing a malicious HuggingFace model with a crafted `quantization_config.quant_dtype` value. When a user loads the model with lmdeploy, the `quant_dtype` is passed to `eval(f'torch.{quant_dtype}')` without any validation. Version 0.12.3 contains a patch.

Published: 2026-09-18T17:13:37.133Z
Updated: 2026-09-22T15:08:07.065Z

InternLM LMDeploy through 0.17.0 contains a reachable assertion vulnerability in the DistServe decode migration loop that allows unauthenticated attackers to terminate the inference engine. Attackers can submit a migration_request with an empty remote_block_ids list to trigger an AssertionError that crashes the engine loop and causes subsequent inference requests to fail.

Published: 2026-09-17T13:43:15.952Z
Updated: 2026-09-17T14:29:19.612Z

InternLM LMDeploy through 0.17.0 in DistServe prefill/decode disaggregation mode fails to release scheduler sessions because the proxy uses user-facing session IDs instead of internal scheduler keys. Unauthenticated attackers can send completion requests to the proxy endpoint that accumulate unreleased scheduler metadata and memory until the prefill worker is out-of-memory killed.

Published: 2026-09-17T14:22:08.232Z
Updated: 2026-09-18T19:18:29.048Z

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Starting in version 0.9.1 and prior to version 0.10.2, the LMdeploy implements an rpc server (AsyncRPCServer in zmq_rpc.py) for supporting the RPC communications. In its core functionality call_and_response(), I found it will directly use the pickles.loads() to deserialize the received messages without any sanitization, hence resulting in a remote code execution vulnerability by this RPC server. Version 0.10.2 contains a patch.

Published: 2026-09-16T15:36:02.095Z
Updated: 2026-09-18T18:05:59.705Z

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Starting in version 0.9.2 and prior to version 0.16.0, LMDeploy's PyTorch DistServe/PD-disaggregation control plane used `recv_pyobj()` to deserialize messages received through a ZeroMQ PULL socket. PyZMQ implements `recv_pyobj()` using Python pickle deserialization, which can execute arbitrary code while reconstructing an object. The peer address used by the receiver was supplied through the `POST /distserve/p2p_connect` HTTP endpoint. An attacker who could reach an affected DistServe API server could cause the server to connect to an attacker-controlled ZeroMQ endpoint and deserialize a crafted pickle payload. API-key authentication is not enabled unless the operator explicitly configures it. As a result, affected DistServe deployments without API keys allowed unauthenticated remote code execution with the privileges of the LMDeploy serving process. This issue affects the PyTorch backend when PD-disaggregation/DistServe is enabled. Ordinary deployments that do not use the affected disaggregated-serving path do not expose this data flow. 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, and block arbitrary outbound ZeroMQ connections from serving nodes. These measures reduce exposure but do not make pickle deserialization safe.

Published: 2026-09-18T17:03:37.084Z
Updated: 2026-09-18T19:54:19.941Z