workspace-cleanup

Workspace cleanup identifies and archives duplicate or obsolete files like `auth.ts`, `auth-new.ts`, and `auth-fixed.ts` using similarity analysis, timestamp checking, and reference scanning. This removes clutter that consumes AI context without losing data—archived files stay retrievable for 30+ days before permanent deletion.

Multi-signal detection scans your directory for temp files, sync conflicts, and abandoned experiments, applying tiered confidence scoring to distinguish clutter from active code. Files move to a central archive with metadata, preserving your project structure while clearing the workspace.

Yes, a two-stage safety workflow archives files first, then prompts for deletion only after 30+ days. This gives you time to verify nothing critical was removed before permanent cleanup occurs.

The Skill supports both recursive cleanup across entire workspace trees and scoped targets for specific directories based on your prompts. You control the scope—clean everything or target particular folders with clutter.

Detection identifies temp files, multiple versions of the same file, sync conflicts, and abandoned experiments through similarity matching, timestamp analysis, and import/reference tracking across your project.