{"id":"CVE-2025-66455","state":"PUBLISHED","title":"LMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.py","description":"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.","publishedAt":"2026-09-18T17:03:37.084Z","updatedAt":"2026-09-18T19:54:19.941Z","severity":{"score":9.8,"label":"Critical","version":"3.1","vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","source":"GitHub_M"},"assigner":"GitHub_M","products":["InternLM lmdeploy"],"affected":[{"vendor":"InternLM","product":"lmdeploy","versions":[{"version":">= 0.9.2, < 0.16.0","status":"affected"}]}],"references":["https://github.com/InternLM/lmdeploy/security/advisories/GHSA-2vh9-42vm-xmv2","https://github.com/InternLM/lmdeploy/commit/f05b4ad8bf2e2d84101a1d63b3c44fadd99223b2","https://github.com/InternLM/lmdeploy/releases/tag/v0.16.0"],"weaknesses":["CWE-502"],"ingestedAt":"2026-10-29T20:09:56.034Z","source":{"repository":"CVEProject/cvelistV5","path":"cves/2025/66xxx/CVE-2025-66455.json","commit":"556c9d3a8cd17e218bb8e452ba7cd619a384ff47"},"raw":{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2025-66455","state":"PUBLISHED","datePublished":"2026-09-18T17:03:37.084Z","dateUpdated":"2026-09-18T19:54:19.941Z","assignerShortName":"GitHub_M","assignerOrgId":"a0819718-46f1-4df5-94e2-005712e83aaa","dateReserved":"2025-12-01T18:44:35.640Z"},"containers":{"cna":{"title":"LMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.py","descriptions":[{"lang":"en","value":"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."}],"affected":[{"vendor":"InternLM","product":"lmdeploy","versions":[{"version":">= 0.9.2, < 0.16.0","status":"affected"}]}],"references":[{"url":"https://github.com/InternLM/lmdeploy/security/advisories/GHSA-2vh9-42vm-xmv2","name":"https://github.com/InternLM/lmdeploy/security/advisories/GHSA-2vh9-42vm-xmv2","tags":["x_refsource_CONFIRM"]},{"url":"https://github.com/InternLM/lmdeploy/commit/f05b4ad8bf2e2d84101a1d63b3c44fadd99223b2","name":"https://github.com/InternLM/lmdeploy/commit/f05b4ad8bf2e2d84101a1d63b3c44fadd99223b2","tags":["x_refsource_MISC"]},{"url":"https://github.com/InternLM/lmdeploy/releases/tag/v0.16.0","name":"https://github.com/InternLM/lmdeploy/releases/tag/v0.16.0","tags":["x_refsource_MISC"]}],"metrics":[{"cvssV3_1":{"attackComplexity":"LOW","attackVector":"NETWORK","availabilityImpact":"HIGH","baseScore":9.8,"baseSeverity":"CRITICAL","confidentialityImpact":"HIGH","integrityImpact":"HIGH","privilegesRequired":"NONE","scope":"UNCHANGED","userInteraction":"NONE","vectorString":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","version":"3.1"}}],"problemTypes":[{"descriptions":[{"cweId":"CWE-502","lang":"en","description":"CWE-502: Deserialization of Untrusted Data","type":"CWE"}]}],"providerMetadata":{"shortName":"GitHub_M","orgId":"a0819718-46f1-4df5-94e2-005712e83aaa","dateUpdated":"2026-09-18T17:03:37.084Z"},"source":{"advisory":"GHSA-2vh9-42vm-xmv2","discovery":"UNKNOWN"}},"adp":[{"title":"CISA ADP Vulnrichment","metrics":[{"other":{"type":"ssvc","content":{"timestamp":"2026-09-18T19:54:08.825899Z","id":"CVE-2025-66455","options":[{"Exploitation":"none"},{"Automatable":"yes"},{"Technical Impact":"total"}],"role":"CISA Coordinator","version":"2.0.3"}}}],"providerMetadata":{"shortName":"CISA-ADP","orgId":"134c704f-9b21-4f2e-91b3-4a467353bcc0","dateUpdated":"2026-09-18T19:54:19.941Z"}}]}}}