clean-project
Skill1 starsChanged 3 months agoMITView on GitHub
File and code clutter reduction; lower cognitive load in the repo and the file explorer. Sweep for temporary artifacts, throwaway verification, obsolete docs, and scratchpad files, then delete or merge them; also strips dead code, unused imports, and duplicate logic, and inlines or folds what one caller uses. Use when a big feature lands, when stabilizing the project, or when asked to "clean up the repo", "remove scaffolding", "reduce clutter", "tidy the repo". Cleaning up the notes or the vault is notes-heartbeat, not this.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- yu-iskw/coding-agent-fabrictheir own version, first seen Feb 5, 2026
- yu-iskw/skill-inspectortheir own version, first seen Jan 23, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- yu-iskw/skill-inspector★ 2same name
- yu-iskw/coding-agent-fabric★ 1same name
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 repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- add-sourceSkill
- analyze-videoSkill
- archive-projectSkill
- code-standardsSkill
- global-heartbeatSkill
- map-projectSkill
- notes-heartbeatSkill
- researchSkill
- resolve-projectSkill
- sleepSkill
- test-projectSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Aggressively prune: Every file, document, test, and hardening has a high maintenance bill.
- Very specific: line 25
Tests: Keep the ones guarding user-visible behavior.
- Very specific: line 26
Docs: Delete non-standard docs on the spot.
- Very specific: line 30
Read the project's priority from $VAULT/projects/<Name>.md before sweeping.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
clean-project
This skill is the sweep after building, for what accumulated: scratch scripts, dead branches, the instrumentation test-project spun up. Same priority posture, later application pass.
Perform an extremely comprehensive and thorough file sweep throughout the entire repo and /private/tmp/claude, applying General rules, File sweep guidelines, and Priority-aware intensity. Aggressively apply changes without hedging, then generate a Closing report.
General rules
- Ignore sunk cost. Age protects nothing. The test: cloning this repo today, would I create this file? If no, delete it.
- Aggressively prune. Every file, document, test, and hardening has a high maintenance bill. Keep it only when the benefit clearly beats that bill. Place any speculative manager, registry, provider, factory, builder, coordinator, adapter, or service under intense scrutiny until a multiple other callers prove otherwise.
File sweep guidelines
For each item, pick whatever reduces long-term complexity: delete, merge, rename, move, inline, split, document, or leave alone. Deletion is the default, not the goal.
- Scratch and temp files. In-repo files (throwaway scripts, one-off verification, migration helpers, experiments, etc), generated artifacts, and session scratchpads at
/private/tmp/claude-<uid>/<cwd-slug>/<session-id>/scratchpad(uid 501 locally, so/private/tmp/claude-501/...).tasks/holds subagent output; clear any stale entries, then judge each artifact worth keeping (produced text/reports, photos, videos, etc). A formal result that stands alongside whatoutput/already holds goes there, named to match the convention already in use; anything one-off, redundant, or unorganized goes tooutput/misc/. - Dead code. Unused imports and functions, obsolete flags, abandoned implementations, unreachable branches, stale assets. Never leave code commented out; git remembers it.
- Duplicate logic. Helpers, parsing, validation, constants, config. Merge when merging clarifies.
- Versioned names. Replace implementations instead of cloning them.
_final,_v2,_new,_old,_copy,_fixed, and_tempare all banned. - Tests. Keep the ones guarding user-visible behavior. Delete the ones whose only job was proving today's implementation ran once. The single runnable check from
code-standardsstays. - Docs. Delete non-standard docs on the spot. Accepted files are
README.mdat the root andtechnical-spec.md,market-analysis.md,masterdoc.html,branch-diff.mdindocs/, perFile structuringincode-standards. Scratch and planning notes belong in the session scratchpad above, never in the repo.
Priority-aware intensity
Read the project's priority from $VAULT/projects/<Name>.md before sweeping. No document defaults to the 1-2 row.
| Priority | Instruction |
| 0 (sandbox, research, hobby) | Aggressive. Minimal files outweigh everything a dead experiment or hardening path might someday be worth. Delete freely. |
| 1-2 (portfolio, research tools) | Standard: temp artifacts, dead code, duplication, version-specific documents. |
| 3 (business, commercial) | Deliberate. Remove genuine cruft, and never strip hardening, error paths, validation, or edge-case handling as "complexity". When torn, justify and retain. |
Closing report
Report what you removed, merged, and refactored, plus anything retained with the reason. The task is done once temporary artifacts and dead code are gone, duplicate logic is merged, regression tests still cover real behavior, and any increase in complexity has a stated and explicit justification.
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
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.