概要
The Hybrid Agent Detector is an architectural audit tool designed to maintain clean separation of concerns within Claude Code projects. It scans for the 'Hybrid Agent' anti-pattern—where orchestration logic and execution tasks are combined in a single agent—which often leads to bloated prompts and reduced code reusability. By analyzing tool usage patterns and file operations, this skill provides a quantitative 'hybrid score' and actionable recommendations for refactoring execution logic into standalone, testable skills.
