CVE-2025-24357
Fixed in a distro release EPSS <1%CVSS 7.5
vLLM is a library for LLM inference and serving. vllm/model_executor/weight_utils.py implements hf_model_weights_iterator to load the model checkpoint, which is downloaded from huggingface. It uses the torch.load function and the weights_only parameter defaults to False. When torch.load loads malicious pickle data, it will execute arbitrary code during unpickling. This vulnerability is fixed in v0.7.0.
Products: vllm-project vllm
Fixed upstream
| Ecosystem | Package | Fixed in |
|---|---|---|
| PyPI | vllm | 0.7.0 |
| PyPI | vllm | d3d6bb13fb62da3234addf6574922a4ec0513d04 |
Distro releases
| Release | Package | Fixed in |
|---|---|---|
| Chainguard | py3.10-vllm-cuda-11.8 | 0.7.2-r0 |
| Chainguard | py3.10-vllm-cuda-12.6 | 0.7.2-r0 |
| Chainguard | py3.10-wheels-vllm-cuda-11.8 | 0.7.2-r0 |
| Chainguard | tritonserver-backend-vllm-24.04 | 24.04-r9 |