agent-sort

To sort repository components, this skill classifies ECC components into DAILY (always loaded) and LIBRARY (searchable) buckets using evidence from your codebase such as file extensions and package managers. It applies parallel subagents to review agents, skills, commands, rules, hooks, and extras efficiently.

Evidence-based ECC management uses your repository's actual codebase as the source of truth to classify components. By analyzing file extensions and package managers, it determines which components belong in the DAILY bucket versus the LIBRARY bucket, reducing noise and improving project efficiency.

Reduce component noise by running this skill to evaluate your repository's codebase and sort Enhanced Code Collection items into DAILY and LIBRARY buckets. This ensures only actively used components are loaded by default, keeping searchable items separate.

Yes, you can classify codebase components in parallel for large repositories. The skill utilizes parallel subagents to perform multiple review passes simultaneously, efficiently categorizing agents, skills, commands, rules, hooks, and extras based on local repository evidence.

Repository optimization for ECC does not require specific package managers but uses them as evidence. The skill analyzes repo-local evidence including file extensions and package managers to make accurate classification decisions for your project-specific components.

The best way to organize ECC components is by sorting them into DAILY and LIBRARY buckets based on codebase evidence. This approach ensures always-loaded components are separated from searchable ones, optimizing your project-specific setup without manual sorting.