CVE-2026-63764
HighCVSS 8.6Exploitation Probability (EPSS)
Low risk23th percentile - higher than 23% of all known CVEs
Summary
SSRF vulnerability in LMDeploy up to version 0.14.0 (fixed in commit 03c3130) in the _load_http_url function in connection.py. The private-IP guard only validates the original URL without re-validating hosts after HTTP redirects.
Risk Assessment
An unauthenticated attacker can use redirects to internal IPs or cloud metadata endpoints, exposing sensitive internal data.
Recommendation
Update LMDeploy to a version containing the fix (commit 03c3130) and consider restricting access to the chat endpoint.
Other vulnerabilities in LMDeploy
See all- CVE-2026-76850Critical
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.
- CVE-2026-46517High
A vulnerability in LMDeploy (versions up to 0.12.3) involves hardcoded "trust_remote_code=True", enabling remote code execution (RCE) via Hugging Face supply-chain attacks without user opt-in.
- CVE-2026-46432High
LMDeploy versions up to 0.12.3 have hardcoded "trust_remote_code=True" in multiple HuggingFace model-loading call sites, allowing arbitrary code execution. No public patches are available at the time of publication.
Original NVD description (English source)
LMDeploy through 0.14.0, fixed in commit 03c3130, contains a server-side request forgery (SSRF) vulnerability in the _load_http_url function within the connection.py media handler, where the private-IP guard validates only the original URL without re-validating hosts after HTTP redirects. An unauthenticated attacker can submit a crafted image_url to the chat completions endpoint pointing to an attacker-controlled host that returns a redirect to a private IP or cloud-metadata endpoint, causing the server to follow the redirect and expose internal service content through the model pipeline.