forgeekscn/claude-custom-skills

Overview

This skill scans and helps clean macOS junk files to free disk space and tidy development projects. It identifies system caches, development build artifacts, duplicate files, and large files, then suggests safe cleanup commands. Use it to reclaim space, speed up the system, and remove unnecessary project artifacts.

How this skill works

The scripts walk target directories and report candidate files for removal: cache and log locations, common development build folders, duplicate files via fdupes, and top large files by size. Each scan produces a reviewable list and recommended shell commands for safe deletion or interactive cleanup. Scripts accept path, depth, and size parameters so you can limit scope and reduce false positives.

When to use it

• Disk space is low and you need quick wins to free space
• Preparing a machine or project for backup or migration
• Cleaning development build artifacts before switching branches or rebuilding
• Finding and removing duplicate media or backup files
• Auditing large files that unexpectedly occupy storage

Best practices

• Always run a scan first and manually review results before deleting
• Limit scan scope with directory, depth, or minimum size to speed up results and avoid mistakes
• Back up important data or export dependency lists before mass deletes
• Use interactive deletion where available (fdupes -d -N) for duplicate groups
• Avoid removing system files you don’t recognize; require sudo only when necessary

Example use cases

• Run scan_cache.sh with sudo to prune user and system caches when macOS feels sluggish
• Use scan_dev_files.sh in a monorepo to find and remove node_modules, .next, and build directories before archiving
• Install fdupes and run scan_duplicates.sh to locate redundant backups in Documents and remove duplicates interactively
• Run scan_large_files.sh to list the top 20 files larger than 1GB and delete or archive the biggest offenders
• Generate targeted cleanup commands and review them before executing to avoid accidental data loss