Storage Cleanup

One-command disk cleanup for macOS and Linux — trash, caches, temp files, old kernels, snap revisions, Homebrew, Docker, and Xcode artifacts. Use when user a...

Security Scan

The skill's code and instructions mostly match a disk-cleanup tool, but the shipped script performs broad destructive operations requiring elevated privileges and the provided file listing/content is truncated so the full behavior could not be verified.

Purpose & Capability

The name, description, SKILL.md and the visible portion of scripts/cleanup.sh align: they scan and remove trash, caches, temp files, package manager caches, journal logs, snap revisions, Homebrew/Xcode artifacts, Docker, etc. The targets the skill touches are consistent with a disk-cleanup utility.

Instruction Scope

SKILL.md instructs users to run the included shell script and documents dry-run and skip flags. The script performs many file-system and package-manager operations (rm -rf, apt clean, brew cleanup, journalctl vacuum, snap remove, sudo find -delete, Docker prunes, etc.). That behavior is in-scope for a cleanup tool, but it's inherently destructive and requires careful review before use. SKILL.md claims 'no dependencies beyond bash and awk' — the script conditionally invokes many external tools (pip, go, brew, apt, snap, docker, journalctl) but checks for their presence; this is reasonable but the claim may be misleading to less-technical users.

Install Mechanism

No install spec — instruction-only with an included script. Nothing is being downloaded or installed by the skill itself, which limits supply-chain risk.

Credentials

No environment variables or external credentials are requested. However, the script invokes sudo for many actions on Linux and runs privileged package-manager and file-deletion commands. Requesting elevated privileges is proportionate to cleaning system-level caches, but it increases risk and requires user caution.

Persistence & Privilege

The skill is not marked always:true and does not request persistent system-wide configuration. It does not declare autonomous-execution restrictions (default behavior), which is expected for skills. There is no evidence it modifies other skills or agent configuration.

What to consider before installing

This skill appears to be a legitimate disk-cleanup utility, but it performs many destructive operations and needs sudo for system-level cleanup. Before installing or running it: 1) Don't run with --yes initially — always run bash scripts/cleanup.sh --dry-run and inspect the results. 2) Review the full scripts/cleanup.sh file yourself (the provided manifest/content was truncated in the review, so confirm the missing tail of the script before trusting it). 3) Back up important data or test in a disposable VM/container first. 4) Pay attention to where it uses sudo and which paths it deletes (ensure no unexpected empty variables or path expansions). 5) If you lack confidence reading shell scripts, ask a knowledgeable admin to audit the complete script. If you want, provide the full, untruncated cleanup.sh so I can re-evaluate with higher confidence.

Reclaim tens of gigabytes in one command. No config files, no dependencies, no damage.

Why This Skill

Systems accumulate junk silently — IDE caches, old snap revisions, stale pip builds, forgotten trash, outdated kernels. Manually hunting them down wastes time and risks deleting the wrong thing.

This skill:

• Scans 12+ cleanup targets across both macOS and Linux in a single pass
• Safe by default — --dry-run shows exactly what would be cleaned before touching anything
• Zero dependencies — pure bash + awk, works on any stock macOS or Linux install
• Cross-platform — auto-detects OS and runs only what applies (no errors on missing tools)
• Selective — skip any category with --skip-kernels, --skip-docker, --skip-brew, --skip-snap
• Reports savings — shows before/after disk usage and exact bytes freed

Quick Start

# Preview what would be cleaned (safe, changes nothing)
bash scripts/cleanup.sh --dry-run

# Clean everything
bash scripts/cleanup.sh --yes

# Clean but keep Docker and old kernels
bash scripts/cleanup.sh --yes --skip-docker --skip-kernels

What Gets Cleaned

Both Platforms

Target	Typical Size	Notes
Trash	1–50 GB	macOS ~/.Trash, Linux ~/.local/share/Trash
Stale /tmp	1–10 GB	pip/npm/rust build dirs older than 60 min
pip cache	50–500 MB	pip cache purge
Go build cache	100 MB–2 GB	go clean -cache
pnpm / yarn / node caches	50–500 MB	Safe to regenerate
JetBrains IDE cache	1–10 GB	IntelliJ, PyCharm, WebStorm, etc.
Whisper model cache	1–5 GB	Redownloads on demand
Chrome / Firefox cache	200 MB–2 GB	Browsing cache only
Playwright browsers	200 MB–1 GB	Redownloads on demand
Docker dangling images	0–10 GB	Only unreferenced images + build cache

Linux Only

Target	Typical Size	Notes
Apt cache	200 MB–2 GB	apt clean
Journal logs	500 MB–4 GB	Vacuumed to 200 MB
Disabled snap revisions	500 MB–5 GB	Old versions kept by snapd
Old kernels	200–800 MB	Keeps current running kernel

macOS Only

Target	Typical Size	Notes
Homebrew old versions	500 MB–5 GB	brew cleanup --prune=7
Xcode DerivedData	2–30 GB	Build artifacts, safe to clear
Xcode Archives	1–20 GB	Old build archives
iOS DeviceSupport	2–15 GB	Old device symbols
CoreSimulator caches	500 MB–5 GB	Simulator disk images
Old user logs	100 MB–1 GB	Logs older than 30 days

Options

Flag	Effect
--dry-run	Preview cleanup without deleting anything
--yes / -y	Run without confirmation prompts
--skip-kernels	Don't remove old kernels (Linux)
--skip-snap	Don't remove disabled snap revisions (Linux)
--skip-docker	Don't prune Docker
--skip-brew	Don't clean Homebrew

Manual Extras

Targets the script doesn't touch (check manually if needed):

• Ollama models: ollama list → ollama rm <unused>
• npm global cache: npm cache clean --force
• Conda envs: conda env list → conda remove -n <env> --all
• Compressed logs: sudo find /var/log -name "*.gz" -delete
• Flatpak (Linux): flatpak uninstall --unused
• Time Machine snapshots (macOS): tmutil listlocalsnapshots / → tmutil deletelocalsnapshots <date>