agentleFS
ohongohong/agent-skills/spring-cleaning/SKILL.md

spring-cleaning

Skill4 starsChanged 2 days agoMITView on GitHub

Inspect Mac disk use, identify recoverable space, and carry out scoped cleanup with verification. Use for low-storage or disk-cleanup requests; keep requests for plans or explanations read-only.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. ohong/skills★ 4100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository10 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the user's report and available local evidence.

    line 15
  2. Distinctive:

    Use it for explicitly named targets or repositories:

    line 24
  3. Distinctive:

    Read category guidance only for relevant categories.

    line 46
  4. Distinctive:

    Read the worktree process before any Git or Codex worktree removal or metadata pruning.

    line 47
  5. Distinctive:

    Prefer supported app removal and package-manager cleanup over deleting internal folders.

    line 62
  6. Distinctive:

    After each useful batch, remeasure the same volume and run the relevant functional check.

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Spring Cleaning

Recover useful disk space and verify the result. A list of suggestions does not complete an authorized cleanup request. Keep plan-only requests read-only. Plan-only requests still permit the read-only inspection needed for a grounded plan.

Inspect a bounded scope

  1. Establish the requested target, such as enough free space for an update. Use the user's report and available local evidence; do not demand a screenshot.
  2. Measure current available space on the affected volume. Inspect likely large targets first, based on the report and actual directories present.
  3. Limit each scan to named paths and a finite timeout. Expand only when results justify it. Avoid a recursive whole-home or root scan by default.
  4. Preserve permission errors, missing volumes, partial scans, and timeouts. An unreadable directory is unknown, not empty. Do not hide stderr.

Resolve the helper from this skill's directory, not the current project directory. Use it for explicitly named targets or repositories:

python3 '/absolute/skill-directory/scripts/inventory.py' --path '/absolute/candidate' --repo '/absolute/repository'

Paths and repositories can repeat. The helper refuses symlink paths, bounds command runtime with a shared budget, and emits JSON. Filesystem metadata calls may block. It never deletes, fetches, or assigns a safe-to-delete verdict. Read its errors and limitations before using results. Directory sizes are estimates of allocated blocks; they are not promised recovery. Parent/child overlap, APFS clones, snapshots, and shared blocks prevent summing them.

Select concrete candidates

Present a short, ranked set of exact paths or app-managed objects. For each, state:

  • Measured size, or why its size is unknown.
  • What it contains and which evidence supports removal or offloading.
  • What must be preserved, restored, or checked before removal.
  • The exact action and a useful verification check.

Read category guidance only for relevant categories. Read the worktree process before any Git or Codex worktree removal or metadata pruning. Clean Git status alone never proves disuse.

Favor large, well-understood candidates with low recovery cost. Do not invent gigabyte estimates, time estimates, or age thresholds that imply safe deletion. Cloud sync is not an independent backup; deletions may propagate to other devices. Confirm actual recoverability for valuable data before proposing its removal.

Execute within the user's scope

Carry existing authorization forward. If exact targets are already authorized, inspect them and proceed without repetitive blanket approval questions. If scope is unresolved, finish the read-only inspection and present concrete targets for the user's decision before deleting them. A screenshot alone authorizes inspection.

Prefer supported app removal and package-manager cleanup over deleting internal folders. Generate commands only for inspected, exact targets. Use quoted paths, no broad globs, and no force flags that bypass a discovered protection. Tool-required flags for a documented, reviewed operation are not blanket force approval. Moving items to Trash is reversible but does not reclaim space until emptied. Empty only the reviewed items within the user's scope; do not clear unrelated Trash.

After each useful batch, remeasure the same volume and run the relevant functional check. Stop when the requested target is met. If no target was given, finish the agreed candidates and report further options without expanding deletion scope.

Close with evidence

Report actions actually completed and the measured change in available bytes. Separate verified recovery from candidate-size estimates and remaining blockers. Concurrent writes and APFS accounting can affect the observed delta; say so when it prevents attributing recovery. State relevant checks actually run and any unverified restoration or application behavior. Do not promise future cleanup.

For macOS storage interpretation, consult Apple's storage guidance. “System Data” is a storage category, not a directory to delete.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.