agentleFS
theseus-runtheseus-run/theseus/.agents/skills/cleanup-audit/SKILL.md

cleanup-audit

Skill1 starsChanged 6 months agoMITView on GitHub

Use after substantial edits, refactors, moves, generated changes, or long agent sessions to find leftovers, stale stubs, unused imports, accidental compatibility artifacts, and trash files. Remove only confirmed non-behavioral leftovers; report anything that may affect compatibility or runtime behavior before changing it.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. About 0 are common, 15 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill as a finishing pass after substantial work.

    line 8
  2. Distinctive:

    Leave the repo without obvious agent leftovers:

    line 12
  3. Distinctive:

    Delete only when the artifact is proven unused, non-public, and non-behavioral.

    line 25
  4. Distinctive:

    If removal might change imports, package exports, runtime behavior, persisted data handling, public contracts, or downstream compatibility, …

    line 27
  5. Distinctive:

    Do not use this skill to redesign code or remove live compatibility paths without confirmation.

    line 31
  6. Distinctive:

    When cleanup collides with compatibility, preserve behavior and report the stale path.

    line 37
  7. Very specific:

    Search old names, moved paths, and temporary terms with rg.

    line 42
  8. Distinctive:

    Check for empty files, stub files, and single-line re-export wrappers.

    line 43
  9. Distinctive:

    Check barrels and package exports for accidental old/new duplication.

    line 44
  10. Distinctive:

    Check imports for old paths, broad namespace artifacts, and unused leftovers.

    line 45
  11. Distinctive:

    Check tests, docs, and comments for stale names introduced by the change.

    line 46
  12. Distinctive:

    Remove only confirmed non-behavioral leftovers.

    line 47
  13. Distinctive:

    test fixture files with no imports and no package export path

    line 59
  14. Distinctive:

    Do not delete without confirmation:

    line 64
  15. Distinctive:

    Use searches shaped by the actual change:

    line 80
  16. Distinctive:

    If nothing was removed, say so clearly and list any concerns found.

    line 96

≈ “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.

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

Cleanup Audit

Use this skill as a finishing pass after substantial work. This is a cleanup and review pass, not a refactor pass.

Goal

Leave the repo without obvious agent leftovers:

  • empty stubs
  • unused imports
  • stale comments
  • accidental re-export wrappers
  • old names in nearby docs or tests
  • duplicate files created during moves
  • scratch artifacts
  • broad import artifacts introduced to avoid updating call sites

Core Rule

Delete only when the artifact is proven unused, non-public, and non-behavioral.

If removal might change imports, package exports, runtime behavior, persisted data handling, public contracts, or downstream compatibility, stop and report it as a concern instead of deleting it.

Not A Refactor

Do not use this skill to redesign code or remove live compatibility paths without confirmation.

  • Cleanup removes confirmed trash.
  • Refactoring changes structure or behavior.
  • Compatibility removal is a breaking change unless the user explicitly authorized it.

When cleanup collides with compatibility, preserve behavior and report the stale path.

Workflow

  1. Inspect git diff --stat and changed files to understand the work surface.
  2. Search old names, moved paths, and temporary terms with rg.
  3. Check for empty files, stub files, and single-line re-export wrappers.
  4. Check barrels and package exports for accidental old/new duplication.
  5. Check imports for old paths, broad namespace artifacts, and unused leftovers.
  6. Check tests, docs, and comments for stale names introduced by the change.
  7. Remove only confirmed non-behavioral leftovers.
  8. Report concerns separately from removed cleanup.

Safe Cleanup Targets

Usually safe when verified unused:

  • empty files that are not imported or exported
  • unused imports
  • scratch files produced by local experiments
  • comments created by the current change that describe obsolete temporary state
  • duplicate local helpers with no callers
  • test fixture files with no imports and no package export path
  • docs or comments that mention an old name where the new name is unambiguous

Red Flags

Do not delete without confirmation:

  • exported symbols
  • package exports
  • public barrels
  • compatibility aliases or shims
  • migration or fallback code
  • persisted-data readers
  • server, RPC, protocol, or serialized contract fields
  • docs archive or design history
  • files still imported anywhere
  • behavior branches with active tests
  • wrappers that may be used by downstream consumers

Search Prompts

Use searches shaped by the actual change:

  • old file stem, new file stem
  • old exported names, new exported names
  • legacy, compat, deprecated, alias, old, v1, temporary, for now, TODO
  • removed path fragments
  • package export paths and barrel names

Report Format

In the final report:

  • Removed: confirmed non-behavioral leftovers.
  • Concerns: possible stale paths that need confirmation before removal.
  • Verification: searches and checks run.

If nothing was removed, say so clearly and list any concerns found.

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.