The Agent Skills Directory

[SAFE]: The skill is designed as a safety layer that enforces a 'human-in-the-loop' requirement. It prohibits any mutating commands or file modifications until the user provides explicit approval through a confirmation gate.
[SAFE]: Tool usage is limited to standard discovery operations such as file reading and directory listing. These tools are used for their intended purposes to gather context and do not involve any network communication or credential access.
[SAFE]: The instructions provide a clear, structured template for summarizing plans to the user, which increases transparency and allows for meaningful review before the agent proceeds with execution.

clarify-confirm-continue