The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests user prompts and the existing project codebase to define the prototype's requirements and structure.
Ingestion points: User prompts and surrounding code modules (SKILL.md).
Boundary markers: Not explicitly defined; the skill relies on the agent's interpretation of the codebase and user request.
Capability inventory: File creation (HTML, JS, TSX) and command execution (pnpm, python, bun, git) as defined in SKILL.md and LOGIC.md.
Sanitization: No explicit sanitization of codebase content is mentioned.
[DYNAMIC_EXECUTION]: The skill is designed to have the agent generate and run executable code locally.
Evidence: LOGIC.md instructs the creation of self-contained HTML files with embedded logic in script blocks. UI.md instructs the creation of React/TSX components for layout variations. This code generation is the primary purpose of the skill.
[COMMAND_EXECUTION]: The skill leverages standard development environment tools to facilitate the execution and storage of prototypes.
Evidence: SKILL.md Rule 2 specifies starting UI prototypes using commands like pnpm <name>, python <path>, or bun <path>. SKILL.md Rule 6 mentions committing prototypes to throwaway git branches.

prototype