hig-doctor

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow is installing and invoking the hig-doctor package from the public GitHub/community source, which at runtime can fetch and ingest outsider-authored repository content (e.g., skill code/templates or any bundled text) into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill instructs running "npx skills add raintree-technology/hig-doctor", which fetches and executes remote package code (linked upstream: https://github.com/raintree-technology/hig-doctor), so installation/runtime relies on external code execution.