CRITICAL9.8PyPI
GHSA-2vh9-42vm-xmv2· CVE-2025-66455, PYSEC-2026-4077LMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.py
Modified: 10/1/2026
package
PyPI/lmdeploy
pkg:pypi/lmdeploy
HIGH7.5PyPI
GHSA-39wr-7q6h-cf68LMDeploy has an SSRF bypass
Modified: 9/18/2026
HIGH8.8PyPI
GHSA-3hmm-rh5q-gwwr· CVE-2026-33625, PYSEC-2026-4078LMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loading
Modified: 10/1/2026
CRITICAL9.8PyPI
GHSA-5h8j-6crg-7rmw· CVE-2025-59953, PYSEC-2026-4079LMdeploy has Remote Code Execution by Pickle Deserialization via zmq_rpc.call_and_response() in InterLM/lmdeploy
Modified: 10/1/2026
HIGH7.5PyPI
GHSA-6w67-hwm5-92mq· CVE-2026-33626, PYSEC-2026-2607LMDeploy has Server-Side Request Forgery (SSRF) via Vision-Language Image Loading
Modified: 7/13/2026
MEDIUM5.3PyPI
GHSA-7vc5-mjwp-c8fq· CVE-2025-3162, PYSEC-2026-1577LMDeploy Improper Input Validation Vulnerability
Modified: 7/7/2026
HIGH8.8PyPI
GHSA-9pf3-7rrr-x5jh· CVE-2025-67729, PYSEC-2026-1578lmdeploy vulnerable to Arbitrary Code Execution via Insecure Deserialization in torch.load()
Modified: 7/7/2026
HIGH7.8PyPI
GHSA-9xq9-36w5-q796· CVE-2026-46517, PYSEC-2026-2608lmdeploy: Hardcoded trust_remote_code=True is an implicit unsafe remote-code load path with no user opt-out
Modified: 9/15/2026
MEDIUM5.3PyPI
GHSA-jfvg-qm4p-473x· CVE-2025-3163, PYSEC-2026-1579InternLM LMDeploy code injection vulnerability
Modified: 7/7/2026
HIGH7.8PyPI
GHSA-m549-qq94-fvhg· CVE-2026-46432, PYSEC-2026-2609LMDeploy: Arbitrary code execution via hardcoded trust_remote_code=True in lmdeploy model initialization
Modified: 7/13/2026
MEDIUM5.3PyPI
PYSEC-2026-1577· CVE-2025-3162, GHSA-7vc5-mjwp-c8fqLMDeploy Improper Input Validation Vulnerability
Modified: 7/7/2026
HIGH8.8PyPI
PYSEC-2026-1578· CVE-2025-67729, GHSA-9pf3-7rrr-x5jhlmdeploy vulnerable to Arbitrary Code Execution via Insecure Deserialization in torch.load()
Modified: 7/7/2026
MEDIUM5.3PyPI
PYSEC-2026-1579· CVE-2025-3163, GHSA-jfvg-qm4p-473xInternLM LMDeploy code injection vulnerability
Modified: 7/7/2026
HIGH7.5PyPI
PYSEC-2026-2607· CVE-2026-33626, GHSA-6w67-hwm5-92mqLMDeploy has Server-Side Request Forgery (SSRF) via Vision-Language Image Loading
Modified: 7/13/2026
HIGH7.8PyPI
PYSEC-2026-2608· CVE-2026-46517, GHSA-9xq9-36w5-q796lmdeploy: Hardcoded trust_remote_code=True is an implicit unsafe remote-code load path with no user opt-out
Modified: 7/13/2026
HIGH7.8PyPI
PYSEC-2026-2609· CVE-2026-46432, GHSA-m549-qq94-fvhgLMDeploy: Arbitrary code execution via hardcoded trust_remote_code=True in lmdeploy model initialization
Modified: 7/13/2026
CRITICAL9.8PyPI
PYSEC-2026-4077· CVE-2025-66455, GHSA-2vh9-42vm-xmv2LMDeploy has Remote Code Execution by Pickle Deserialization via handle_zmq_recv in lmdeploy/lmdeploy/pytorch/disagg/conn/engine_conn.py
Modified: 10/1/2026
HIGH8.8PyPI
PYSEC-2026-4078· CVE-2026-33625, GHSA-3hmm-rh5q-gwwrLMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loading
Modified: 10/1/2026
CRITICAL9.8PyPI
PYSEC-2026-4079· CVE-2025-59953, GHSA-5h8j-6crg-7rmwLMdeploy has Remote Code Execution by Pickle Deserialization via zmq_rpc.call_and_response() in InterLM/lmdeploy
Modified: 10/1/2026