vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.6.5 and prior to 0.8.5, having vLLM integration with mooncake, are vulnerable to remote code execution due to using pickle based serialization over unsecured ZeroMQ sockets. The vulnerable sockets were set to listen on all network interfaces, increasing the likelihood that an attacker is able to reach the vulnerable ZeroMQ sockets to carry out an attack. vLLM instances that do not make use of the mooncake integration are not vulnerable. This issue has been patched in version 0.8.5.
History

Wed, 30 Apr 2025 00:45:00 +0000

Type Values Removed Values Added
Description vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.6.5 and prior to 0.8.5, having vLLM integration with mooncake, are vulnerable to remote code execution due to using pickle based serialization over unsecured ZeroMQ sockets. The vulnerable sockets were set to listen on all network interfaces, increasing the likelihood that an attacker is able to reach the vulnerable ZeroMQ sockets to carry out an attack. vLLM instances that do not make use of the mooncake integration are not vulnerable. This issue has been patched in version 0.8.5.
Title vLLM Vulnerable to Remote Code Execution via Mooncake Integration
Weaknesses CWE-502
References
Metrics cvssV3_1

{'score': 10, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H'}


cve-icon MITRE

Status: PUBLISHED

Assigner: GitHub_M

Published: 2025-04-30T00:25:00.655Z

Updated: 2025-04-30T00:25:00.655Z

Reserved: 2025-04-08T10:54:58.369Z

Link: CVE-2025-32444

cve-icon Vulnrichment

No data.

cve-icon NVD

Status : Received

Published: 2025-04-30T01:15:51.953

Modified: 2025-04-30T01:15:51.953

Link: CVE-2025-32444

cve-icon Redhat

No data.