agentleFS
RektTillNoonRektTillNoon/alpha-skills/owner-clean/SKILL.md

owner-clean

Skill4 starsChanged 6 months agoMITView on GitHub

Post-change ownership and cleanup pass for Codex's own work. Use when the user asks to run owner-check and clean on your work, audit recent edits before finishing, patch bugs found during verification, make a completed change leaner without product drift, or ensure an implementation is owned at the right seam before reporting done.

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. About 0 are common, 13 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈13

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run a finishing pass over the current work: verify the intended edit is owned at the right layer, perform one bounded behavior-preserving …

    line 10
  2. Distinctive:

    Run one ownership lens.

    line 21
  3. Very specific:

    Use owner-check before cleanup when the work touches canonical paths, duplicated state, lifecycle authority, cache/sync/reconnect behavior, …

    line 22
  4. Distinctive:

    Keep it non-recursive: perform one owner-check pass, record the owner decision, then resume this same owner-clean pass.

    line 23
  5. Distinctive:

    If ownership is not relevant, state that as an assumption and continue to cleanup.

    line 24
  6. Very specific:

    Use clean for the cleanup discipline: smallest behavior-preserving edits, narrow scope, nearby tests, and no product expansion.

    line 27
  7. Distinctive:

    Prefer deleting false owners, stale aliases, duplicate paths, misleading names, brittle tests, and low-value indirection found during the …

    line 28
  8. Distinctive:

    Stop after the highest-signal 1-3 improvements unless verification exposes directly related bugs.

    line 29
  9. Distinctive:

    Treat failing tests, type errors, broken docs checks, invalid generated artifacts, or owner-invariant violations as part of the current …

    line 32
  10. Distinctive:

    Do not broaden into unrelated failures.

    line 33
  11. Distinctive:

    Verify with the cheapest sufficient proof.

    line 35
  12. Distinctive:

    Prefer the narrowest relevant test, typecheck, lint, build, screenshot, link check, or static reference search that proves the touched …

    line 36
  13. Distinctive:

    Run broader gates only when shared behavior changed, local instructions require them, or the user asked.

    line 37
  14. Distinctive:

    Use this compact shape when useful:

    line 41
  15. Distinctive:

    Keep the report direct. Include commands actually run and failures actually observed.

    line 53

≈ “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

Owner Clean

Goal

Run a finishing pass over the current work: verify the intended edit is owned at the right layer, perform one bounded behavior-preserving cleanup pass, patch bugs found by that process, and report the proof.

This skill composes $owner-check and $clean; it does not replace either skill or copy their full workflows. Read and follow those installed skills when they are available.

Workflow

  1. Discover the active work surface.
    • Inspect the current branch or working tree when there is a repo.
    • Identify the files, tests, generated artifacts, and user-requested scope that belong to the current task.
    • Preserve unrelated user edits.
  2. Run one ownership lens.
    • Use owner-check before cleanup when the work touches canonical paths, duplicated state, lifecycle authority, cache/sync/reconnect behavior, render authority, or semantic ownership.
    • Keep it non-recursive: perform one owner-check pass, record the owner decision, then resume this same owner-clean pass.
    • If ownership is not relevant, state that as an assumption and continue to cleanup.
  3. Run bounded cleanup.
    • Use clean for the cleanup discipline: smallest behavior-preserving edits, narrow scope, nearby tests, and no product expansion.
    • Prefer deleting false owners, stale aliases, duplicate paths, misleading names, brittle tests, and low-value indirection found during the owner lens.
    • Stop after the highest-signal 1-3 improvements unless verification exposes directly related bugs.
  4. Patch bugs found in the process.
    • Treat failing tests, type errors, broken docs checks, invalid generated artifacts, or owner-invariant violations as part of the current pass when they are caused by the current work.
    • Do not broaden into unrelated failures. Separate observation, inference, and assumption before deciding a failure belongs to the current work.
  5. Verify with the cheapest sufficient proof.
    • Prefer the narrowest relevant test, typecheck, lint, build, screenshot, link check, or static reference search that proves the touched contract.
    • Run broader gates only when shared behavior changed, local instructions require them, or the user asked.

Reporting

Use this compact shape when useful:

Observations:
Inferences:
Assumptions:
Owner:
Cleanup:
Proof:
Left out:

Keep the report direct. Include commands actually run and failures actually observed. Do not claim a clean result unless verification ran or the remaining risk is explicitly stated.

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.