onnx-webgpu-converter

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill fetches and inspects public, user-uploaded model repos and metadata from the HuggingFace Hub (e.g., browsing huggingface.co, referencing onnx-community models, and the scripts/preflight_check.py calls to huggingface_hub.model_info and transformers.AutoConfig.from_pretrained), so it ingests untrusted third‑party content that the agent reads to drive its workflow.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains an explicit runtime CDN import that fetches and executes remote JavaScript — import { pipeline } from "https://cdn.jsdelivr.net/npm/@huggingface/transformers@3" — which is a runtime external dependency that executes remote code.