The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to verify the generated prototype in Phase 3.1. It runs scripts defined in the repository's package.json (such as 'lint', 'dev', 'start', or 'serve') and the TypeScript compiler ('tsc'). These commands are standard for frontend development environments but involve executing scripts defined in the user's project.- [PROMPT_INJECTION]: The skill processes untrusted data from PRDs and User Journeys to generate code and mock data, creating an indirect prompt injection surface. Ingestion points: PRD and User Journey documents read in Phase 1.1. Boundary markers: None specified to delimit ingested text from instructions. Capability inventory: File-write access (sandbox directory) and command execution via shell scripts (npm/tsc). Sanitization: No explicit sanitization of document content is described.- [DATA_EXFILTRATION]: The skill scans the project directory and reads configuration files (e.g., package.json, tsconfig.json, vite.config.ts) to identify the tech stack. It implements safeguards by excluding sensitive directories like .git/, .next/, and build/ from scans.

prototype-designer