agentleFS
kuya-eggkuya-egg/Monozukuri/kaizen/SKILL.md

kaizen

Skill11 starsChanged 2 months agoMITView on GitHub

Improve software continuously through small, safe, behavior-preserving changes that reduce maintenance cost. Use for refactoring, cleanup, technical debt, naming, duplication, dead code, or post-change polish.

18 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 12 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 12

    Distinctive

    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:

    does not turn a focused task into a rewrite.

    line 30
  2. Distinctive:

    Run the narrowest relevant tests, build, type check, benchmark, or behavior observation before changing structure.

    line 42
  3. Distinctive:

    Run the baseline checks and inspect the diff.

    line 58
  4. Distinctive:

    If more work is needed, record a concrete follow-up with a reason and boundary.

    line 63
  5. Distinctive:

    Do not create a vague technical-debt backlog as a substitute for judgment.

    line 64
  6. Distinctive:

    Do not refactor unrelated code because it is aesthetically unpleasant.

    line 74
  7. Distinctive:

    Do not rewrite a stable module without a concrete problem and migration path.

    line 75
  8. Distinctive:

    Do not change behavior silently under the label of cleanup.

    line 76
  9. Distinctive:

    Do not add abstraction unless it removes more complexity than it creates.

    line 77
  10. Distinctive:

    Do not use “leave it better” to bypass the user's scope.

    line 78
  11. Distinctive:

    Do not reformat untouched code or apply broad lint fixes under the name of Kaizen.

    line 79
  12. Distinctive:

    If a referenced skill is not installed, apply its named lens inline instead of trying to invoke it.

    line 88

“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “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

Kaizen

Kaizen is continuous improvement. In software, make the codebase a little clearer, safer, or easier to change whenever a relevant opportunity can be handled without hiding scope or risk. This is a Japanese-inspired engineering metaphor, not permission for constant churn.

Use this skill when

  • refactoring an existing area;
  • removing duplication, dead paths, or misleading names;
  • improving a module while implementing adjacent behavior;
  • reducing recurring maintenance cost;
  • converting a lesson into a guardrail; or
  • deciding whether cleanup belongs in the current change.

Outcome

Produce a small improvement that:

  • has a concrete maintenance, correctness, safety, or clarity benefit;
  • preserves behavior unless behavior change is explicit;
  • is easy to review and revert;
  • has appropriate regression protection; and
  • does not turn a focused task into a rewrite.

Workflow

1. Name the waste

Identify the specific pain: misleading language, duplication, dead code, unclear boundary, missing test, noisy diagnostic, fragile setup, repeated manual step, or known defect. Explain why it matters now.

2. Establish a baseline

Run the narrowest relevant tests, build, type check, benchmark, or behavior observation before changing structure. If no baseline exists, state the risk.

3. Make one small move

Rename, extract, delete, simplify, constrain, document, or automate one clear improvement. Keep the change local and avoid mixing a behavior change with an unrelated cleanup unless the cleanup is required for safety.

An adjacent improvement is allowed only when it is directly related to the task, low risk, independently verifiable, smaller than the primary change, and unlikely to complicate review or rollback. Otherwise record it as a concrete follow-up.

4. Preserve and verify

Run the baseline checks and inspect the diff. Confirm that the improvement reduced concepts or risk rather than merely moving complexity elsewhere.

5. Record the next step only when real

If more work is needed, record a concrete follow-up with a reason and boundary. Do not create a vague technical-debt backlog as a substitute for judgment.

Evidence standard

Show the before-and-after reason the change improves maintainability or safety. For performance or reliability claims, measure or state that the claim is not measured.

Boundaries

  • Do not refactor unrelated code because it is aesthetically unpleasant.
  • Do not rewrite a stable module without a concrete problem and migration path.
  • Do not change behavior silently under the label of cleanup.
  • Do not add abstraction unless it removes more complexity than it creates.
  • Do not use “leave it better” to bypass the user's scope.
  • Do not reformat untouched code or apply broad lint fixes under the name of Kaizen.

Handoff

Report the maintenance problem, improvement, behavior checks, and remaining opportunities. Use poka-yoke for a new guardrail, kodawari for review, or hansei when the improvement came from a failure or incident.

If a referenced skill is not installed, apply its named lens inline instead of trying to invoke it.

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.