- Home
- Skills
- Productivity
- Dev Disk Reclaimer
Works with the AI Agents you already use
Dev Disk Reclaimer
Scan macOS and Linux development caches and get a risk-labelled reclaim plan without deleting anything.
$5
Dev Disk Reclaimer
Example session with this skill installed
My Mac is out of space. What dev caches can I clear safely?
- Read your context and instructions
- Compiled the dev disk reclaimer
Disk reclaim report — macOS
Free: 18.2 GB / 494 GB
| Risk | Location | Size | Last used | Regenerates |
|---|---|---|---|---|
| safe | ~/.npm/_cacache | 6.4 GB | 2 days ago | next npm install |
| safe | ~/Library/Caches/Homebrew | 3.1 GB | 1 week ago | next brew install |
| safe | ~/.cargo/registry | 2.0 GB | 3 days ago | next cargo fetch |
| caution | ~/Library/Developer/Xcode/DerivedData | 12.8 GB | today | next Xcode build |
| caution | ~/Library/Containers/com.docker.docker/Data | 9.4 GB | today | Docker manages this |
| never | ~/Library/Application Support/Postgres | — | not scanned | user data |
Safe: 11.5 GB · Caution: 22.2 GB
Never-touch paths are listed but were not scanned.
Self-check
| Check | Result |
|---|---|
| no duplicate paths | pass |
| all never-touch paths reported | pass |
| never-touch paths not scanned | pass |
| safe/caution entries measured | pass |
| risk labels valid | pass |
| all report paths under home | pass |
| free space <= total space | pass |
| no skipped path counted | pass |
PASS — the report is safe to act on.
Nothing was deleted. Commands for the safe tier:
npm cache clean --force
brew cleanup --prune=all
then re-run the scan to see the delta
Start with the safe tier. Review the Docker volumes before touching anything
in the caution tier: docker volume ls, never prune --volumes blindly.
Connects securely to your tools. The creator never sees your data.
About this skill
Developer machines fill up in places nobody looks: package manager caches, Xcode DerivedData, Docker's data store, old build outputs. The same happens on CI runners and Linux dev servers, where nobody is watching the disk either. Deleting the wrong thing costs an afternoon; deleting nothing costs the disk.
What it does
- Cache scan across the common development categories, from npm and Cargo to Xcode, Docker, Gradle, and browser caches, on macOS and Linux.
- Risk tiers that separate safe, caution, and never-touch locations.
- Scope guards that never follow nested symlinks, never cross into another filesystem, skip anything outside your home, and report unreadable paths instead of dropping them.
- Read-only operation that never deletes, moves, or runs
rm. - Cleanup plan with ordered commands, expected regeneration, and a re-scan to verify.
How it works
- Detect the OS (
macOSorLinux) and load the matching cache map. - Scan the known cache roots and measure size, file count, and last use.
- Classify each item by risk, skipping database directories entirely.
- Present the largest items first with the exact command the user may run.
- Re-scan after cleanup and compare the freed space.
Frameworks & tools
Python 3.9+ on macOS or Linux. No sudo, no network, and no writes except the report files you name.
Why this beats prompting it yourself
Prompts happily suggest rm -rf on a database directory or docker prune --volumes without warning. This skill never deletes, never scans user data or anything outside your home, labels every risk, and fails closed if its own safety check trips.
- Covers macOS and Linux; Windows is not covered yet.
- Sizes use allocated blocks, not
du-exact; hardlinked files may be counted per path. - Caches that resolve outside the home directory are reported as skipped, not measured.
- Does not delete anything; you run the commands.
- Docker volume cleanup needs its own judgement and is marked caution.
- New tools with unusual cache locations may not be in the default list.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 26 days ago
Requires
Permissions
File Scopes