The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project documentation (PRD, Tech Spec, Tasks) and browser content (DOM snapshots, console logs) without explicit sanitization or boundary markers. Maliciously crafted instructions inside these external sources could attempt to influence the agent's behavior or reporting during the QA process.
Ingestion points: Files located at ./tasks/prd-[feature-slug]/prd.md, techspec.md, tasks.md, and application content via Playwright MCP tools.
Boundary markers: Absent; the skill does not define delimiters for untrusted content.
Capability inventory: File reading, report generation (file writing), and full browser interaction (navigation, clicking, typing, key simulation).
Sanitization: No explicit sanitization or filtering of external data is defined.
[DATA_EXPOSURE]: The skill is designed to read sensitive project documentation, including functional requirements and technical specifications from the local filesystem. This is consistent with its primary purpose for QA validation.
[COMMAND_EXECUTION]: The instructions include guidance for the user to run development commands (bun run dev, bun run test:e2e) if the environment is not set up correctly. The agent does not execute these commands itself.

executar-qa