The Agent Skills Directory

[SAFE]: The skill performs context gathering by reading local files such as .claude/docs/technical-preferences.md and checking for the presence of directories like src/ and prototypes/ to determine the project's progress.
[SAFE]: It uses interactive prompts to guide the user through a decision tree, recommending appropriate next steps based on user input.
[SAFE]: The skill writes a simple configuration value to production/review-mode.txt based on a user's choice, which is standard behavior for setup tools.
[SAFE]: No network access, remote code downloads, or suspicious data operations were identified.

start