The Agent Skills Directory

[SAFE]: The skill establishes a strict operational boundary by stating that the agent is not allowed to implement, code, or modify behavior while the brainstorming skill is active. This serves as a primary defense against unauthorized or accidental system modifications.
[SAFE]: The 'Understanding Lock' mechanism acts as a mandatory human-in-the-loop checkpoint, requiring explicit user confirmation of the project's summary and assumptions before the agent can proceed to the design phase.
[SAFE]: The skill proactively addresses security by making the discussion of non-functional requirements (including security and privacy constraints) a mandatory part of the design process.
[SAFE]: The process requires incremental validation and clear documentation through a Decision Log, ensuring all design choices are transparent and reviewed rather than hidden in automated actions.

brainstorming