The Agent Skills Directory

[SAFE]: The skill establishes secure practices for prototyping by enforcing code isolation and minimal data persistence.
Evidence: Rule 3 in SKILL.md states that state should live in memory by default and scratch databases should be clearly marked as prototypes.
[SAFE]: Terminal interaction instructions are confined to standard ANSI escape sequences for formatting and display management.
Evidence: LOGIC.md suggests using standard codes like \x1b[1m for bolding and \033[2J for clearing the screen.
[SAFE]: UI prototyping guidelines include environment-based security gates to prevent debug tools from reaching production.
Evidence: UI.md instructs the agent to gate the prototype switcher component on process.env.NODE_ENV !== 'production'.
[SAFE]: The skill focuses on project hygiene and temporary code lifecycle.
Evidence: Rule 6 in SKILL.md and Step 6 in UI.md explicitly instruct the agent to delete or merge the prototype code once it has served its purpose.

prototype