CVE-2026-76850Critical· 9.8▾ AbyssalPoC availableLMDeploy deserializes disaggregated-serving peer messages with pickle. The handle_zmq_recv coroutine in lmdeploy/pytorch/disagg/conn/engine_conn.py reads peer-to-peer cache-free requests with recv_pyobj(), which deserializes the received…
▾ Abyssal zone — Critical with a public exploit or in-the-wild use
impact 53.9 · likelihood 0.2 · exploitation 12
A public proof-of-concept already exists for this vulnerability — see Exploit availability below.
Stakeholder-Specific Vulnerability Categorization from CISA's ADP record at CVE.org: whether exploitation is observed, whether an attack can be automated, and how much of the system is at stake. The CVSS score shown above comes from the assigning CNA record, not NVD.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via CVEORG
Last analysed / modified upstream
1.0%
LMDeploy deserializes disaggregated-serving peer messages with pickle. The handle_zmq_recv coroutine in lmdeploy/pytorch/disagg/conn/engine_conn.py reads peer-to-peer cache-free requests with recv_pyobj(), which deserializes the received bytes with pickle.loads(), and the isinstance check against DistServeCacheFreeRequest runs only after deserialization has already completed. The peer that supplies those bytes is caller-controlled: p2p_connect passes remote_engine_endpoint_info.zmq_address from the request body to connect() on the ZMQ PULL socket, and the POST /distserve/p2p_initialize and /distserve/p2p_connect endpoints in lmdeploy/serve/openai/api_server.py apply no authentication unless the server is started with api_keys, which defaults to None. A remote attacker can direct an engine to pull from a ZMQ endpoint under their control and execute arbitrary code in the engine process. Deployments that do not enable disaggregated serving are not affected, because the receive loop is only started once the migration backend accepts the connection.
lmdeploy >= 0.9.2 < 0.16.0Refer to the linked advisories for vendor-supplied fixes and affected version ranges.
Connected by shared product, vendor, weakness, or advisory.
CVE-2025-66455Critical· 9.8LMDeploy is a toolkit for compressing, deploying, and serving large language models
CVE-2025-59953Critical· 9.8LMDeploy is a toolkit for compressing, deploying, and serving large language models
CVE-2026-33625High· 8.8LMDeploy is a toolkit for compressing, deploying, and serving large language models
CVE-2026-92983High· 7.5InternLM 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
CVE-2026-92971High· 7.5InternLM 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
CVE-2026-33626High· 7.5LMDeploy has Server-Side Request Forgery (SSRF) via Vision-Language Image Loading