The Agent Skills Directory

[SAFE]: The skill implements a robust human-in-the-loop pattern, using the AskUserQuestion tool to require user approval at every phase transition before proceeding.
[SAFE]: File operations and complex tasks are delegated to subagents using the Task tool, following a structured workflow that limits the orchestrator's direct execution of high-risk commands.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting project documentation and user arguments to parameterize subagents. Ingestion points: Feature arguments, design/gdd/ files, sound-bible.md, and technical-preferences.md. Boundary markers: None explicitly defined in the orchestration logic. Capability inventory: The orchestrator and its subagents have access to Bash, Write, Edit, and Task tools. Sanitization: No input validation or sanitization is performed on the ingested project data. This risk is mitigated by the mandatory user approval gates.

team-audio