vLLM是vLLM开源的一个适用于 LLM 的高吞吐量和内存高效推理和服务引擎。 vLLM 0.6.5至0.8.5之前版本存在代码问题漏洞,该漏洞源于使用基于pickle的序列化,可能导致远程代码执行。
Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.
10-question deep dive: root cause, exploitation, mitigation, urgency. Read summary free, full version requires login.
| Vendor | Product | Affected Versions | CPE | Subscribe |
|---|---|---|---|---|
| vllm-project | vllm | >= 0.6.5, < 0.8.5 | - |
|
| # | POC Description | Source Link | Shenlong Link |
|---|---|---|---|
| 1 | None | https://github.com/stuxbench/vLLM-CVE-2025-32444 | POC Details |
| 2 | None | https://github.com/stuxbench/vllm-cve-2025-32444 | POC Details |
| CVE-2025-30202 | 7.5 HIGH | Data exposure via ZeroMQ on multi-node vLLM deployment |
| CVE-2025-46560 | 6.5 MEDIUM | vLLM phi4mm: Quadratic Time Complexity in Input Token Processing leads to denial of servi |
No comments yet