outlines

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill includes runtime model loading from a Hugging Face repo (e.g. outlines.models.vllm("meta-llama/Llama-3.1-8B-Instruct", trust_remote_code=True)), which implies fetching and running remote repository code from https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct at runtime and therefore can execute remote code.