clone-prototype

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

• Third-party content exposure detected (low risk: 0.10). The workflow consumes screen captures and references supplied by the user, and optionally uses the Mobbin MCP tool (mcp__mobbin__search_screens, search_flows) which returns search results and remote image URLs that are subsequently downloaded via curl. Since the agent actively fetches specific outsider-authored content or search results during the workflow, the exposure risk is low.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill references an external repository at https://github.com/ReScienceLab/super-prototyping#subdirectory=tools installed via uv tool install during setup, constituting an unverifiable runtime/setup dependency.