lingbot-map-3d-reconstruction

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs fetching and ingesting third-party artifacts at runtime (git clone from GitHub, model checkpoints via huggingface_hub/huggingface.co and ModelScope, wget of skyseg.onnx from huggingface, and pip installs from external indexes like flashinfer.ai), which are untrusted/user-provided resources consumed by the skill and could materially alter runtime behavior.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill's installation fetches and installs remote code (git clone https://github.com/Robbyant/lingbot-map.git followed by pip install -e .), which downloads and executes third-party code during setup and is therefore a runtime/execution risk.