概要
The edit-agent skill streamlines the development of sub-agents within the Claude Code environment by automating the setup of YAML frontmatter, tool permissions, and system prompt architectures. It acts as an intelligent orchestrator that evaluates whether a task truly requires a full sub-agent or could be handled by a simpler tool like a skill or bash script, ensuring efficient token usage and context isolation. By enforcing single-responsibility principles and minimal tool access, it helps developers build a modular and secure multi-agent ecosystem for complex, multi-step AI reasoning tasks.
