agentleFS
otaviobarrosotaviobarros/agent-engineering/skills/implement-change/SKILL.md

implement-change

Skill3 starsChanged 20 days agoMITView on GitHub

Implement an accepted software change plan with scoped code, tests, and documentation. Use for the implementation phase or to resume an authorized slice. Standalone implementation stops before independent review and publication.

17 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use audit-test-suite, change-data-safely and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • yaohuangguanSkill

    audit-test-suite

    Find where an existing test suite is weak and decide what to test next, ranked by risk. Use this when asked to add missing tests, find gaps, raise coverage, assess test quality or see which areas are thin, even if the request only mentions a coverage percentage.

    Also in 3 projects that use this

  • yaohuangguanSkill

    change-data-safely

    Change a database schema or stored data without breaking running code or losing data. Use this for any migration, new or changed column, backfill, bulk update, delete, index, or change to a stored format, including in small projects. Data mistakes are often impossible to undo.

    Also in 3 projects that use this

  • SauravSrivastavSkill

    debug-failure

    Diagnose a failing test, command, or runtime behavior using evidence and produce a minimal verified fix.

    Also in 3 projects that use this

  • MicrosoftSkill

    fix-flaky-test

    Reproduces and fixes flaky or quarantined tests. Tries local reproduction first (fast), then falls back to CI reproduce workflow (reproduce-flaky-tests.yml). Use this when asked to investigate, reproduce, debug, or fix a flaky test, a quarantined test, or an intermittently failing test.

    Also in 3 projects that use this

  • srclightSkill

    refactor-safely

    This skill should be used when the user asks to "refactor", "rename", "move", "change a function signature", "extract a method", "delete unused code", "clean up", or any modification that could break callers or dependents. Provides a pre-change safety checklist using srclight's graph tools to assess blast radius before touching code.

    Also in 5 projects that use this

  • +6 more from yaohuangguan

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 7

    Distinctive

    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:

    After reading affected code, follow this reuse-first ladder.

    line 29
  2. Distinctive:

    Search the codebase for an existing helper or pattern before writing a helper, parser, formatter, or validator.

    line 33
  3. Distinctive:

    Use a standard library or platform feature.

    line 36
  4. Distinctive:

    Use an installed dependency.

    line 37
  5. Distinctive:

    Never cut validation at trust boundaries, error handling that prevents data loss, security, or accessibility.

    line 40
  6. Distinctive:

    Run relevant development checks within authority and record command, candidate, outcome, and limitation.

    line 57
  7. Distinctive:

    After the last edit and check, compare the final diff, staged and untracked contents, branch, and revision with the slice and existing work.

    line 68
  8. Very specific:

    If the task mentions speculative or future needs, include a Not built: … line naming what you left out.

    line 82

“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

Implement change

Produce the authorized changes and usable verification inputs. First read the contract core and the verification rules.

1. Confirm the plan and authority

  • Identify the requirements owner, accepted plan, criteria, authorized slice, and limits. Confirm workspace, branch, revision, and existing staged and untracked changes.
  • Revalidate the plan against current inputs and reuse its valid decisions. Open criterion or material decision → report it and keep dependent work pending.
  • An accepted plan alone does not authorize edits.

Checkpoint 1: the slice has an accepted expected result and implementation authority, or a clear blocker.

2. Ground the slice

  • Trace affected entrypoints, callers, contracts, test procedures, and executable check configuration. Follow project conventions and existing interfaces; use the plan's checks. Material redesign or scope expansion → get a decision.
  • After reading affected code, follow this reuse-first ladder. Take the first option that fully meets the accepted criteria:

    1. Skip what the request does not need, including unrequested features and speculative options; name what you skip.
    2. Search the codebase for an existing helper or pattern before writing a helper, parser, formatter, or validator. State what you found or did not find; reuse it the way surrounding code does.
    3. Use a standard library or platform feature.
    4. Use an installed dependency. Never add one for a few lines.
    5. Otherwise, write the minimum code that works.

    Never cut validation at trust boundaries, error handling that prevents data loss, security, or accessibility. New non-trivial logic still needs a test.

  • Bug fix → capture original-failure evidence before corrective edits when available; otherwise state the limitation.

Checkpoint 2: edits and test expectations rest on inspected code and accepted criteria. Pre-existing failures are recorded separately.

3. Implement

  • Make scoped code, test, and documentation changes; keep accepted behavior outside the slice. Apply the verification rules. A bug regression exercises the interface where the original failure occurred.
  • Reuse project test mechanisms. Propose missing infrastructure and its scope before adding it within authority; beyond scope it is a required gap.
  • Dependency installation, migrations, external effects, and experiments need their own authority; the plan does not expand execution permission.
  • Run relevant development checks within authority and record command, candidate, outcome, and limitation. Keep evidence before cleaning up resources you created.
  • Fix scoped defects within the three-round limit in the review rules. A blocked required check leaves unaffected work free to continue.

Checkpoint 3: authorized edits, tests, and docs are present, or incomplete items are explicit. A passing local check is not full verification.

4. Inspect and hand off

After the last edit and check, compare the final diff, staged and untracked contents, branch, and revision with the slice and existing work. Run the contract's freshness check for earlier evidence. Then hand off in an authorized record or the response:

Behavior implemented:
Criteria affected: (reference the plan)
Checks: command | candidate | result | limitation
Not verified: criteria whose required check did not run, and why
Final inspection: branch, revision, dirty inputs, scope and preservation result
Remaining work, gaps, next action:

If the task mentions speculative or future needs, include a Not built: … line naming what you left out.

Carry Not verified into any status the user asks for.

Incomplete inspection → name the missing evidence and mark the handoff incomplete.

Done when changes and verification inputs are accounted for, with failures and gaps visible, and every criterion whose required check did not run is listed under Not verified in the handoff and any requested status. Implementation completion is not verification, review, or delivery readiness. A standalone request stops here; authorized delivery continues with verify-change and review-change. No commits, pushes, or publication.

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.