agentleFS
Sign inStart free
taljeontaljeon/plzdo/skills/implement-change/SKILL.md

implement-change

Skill9 starsChanged 16 days agoMITView on GitHub

Claude writes the spec, the implementer agent builds and tests it in a git worktree, review-change reviews it, and Claude gives final acceptance. Use for approved multi-file features, new modules, or refactors. Skip for small single-file edits (do them directly) and for coding-test practice (coding-test-coach).

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 10 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

  • BanKnightSkill

    plan-change

    将已完成 spec/design 的 change 拆成可执行 plan 与 tasks。用户要把设计落成执行计划、任务清单或实现步骤时使用。

    Also in 6 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

5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 3

    Distinctive

    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:

    Commit the work the implementer needs (ask the user if committing is not already approved).

    line 14
  2. Very specific:

    Start the implementer agent with the packet.

    line 36
  3. Very specific:

    Run the review-change skill on the implementer's branch against the base commit, with the spec and the implementer's reply.

    line 40
  4. Distinctive:

    Read the diff against the spec, requirement by requirement.

    line 45
  5. Distinctive:

    Check out the branch, rerun the acceptance command yourself, and read the exit code.

    line 46

“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

Spec -> implementer -> review -> acceptance

Claude-only version of the owner's codex-implement flow. Claude owns the design, the spec and the final acceptance; the implementer agent writes the code in its own worktree.

1. Record the base

A worktree starts from a commit and never sees uncommitted work.

  1. Commit the work the implementer needs (ask the user if committing is not already approved).
  2. Record git rev-parse HEAD as the base commit, and git status --porcelain so you can tell later which changes are yours.

2. Write the spec packet

The implementer cannot see this conversation, so the packet must stand alone. Use the design-spec output when there is one.

You are the implementation worker. Follow this spec exactly.
Base commit: <sha>. Branch: <name>.
Objective: <one line>
Requirements:
1. <behavior, with edge cases and error types>
2. ...
Allowed paths: <paths>. Do not touch anything else.
Tests: reuse existing checks; test each required behavior (normal, edge, error) at the interface named above, not internals.
Acceptance: run '<command>' from the repository root; it must exit 0.
Reply with: branch, diffstat, the exact test summary line you observed, the acceptance exit code, status (completed | partial | unknown | not_started), and any spec ambiguity you resolved.

3. Run

Start the implementer agent with the packet. At most 3 implementers in parallel, and their allowed paths must not overlap. Wait for every one to finish before reporting.

4. Review

Run the review-change skill on the implementer's branch against the base commit, with the spec and the implementer's reply. Leave out your own verdict.

5. Final acceptance

  1. git diff --stat <base>..<branch>: only allowed paths changed.
  2. Read the diff against the spec, requirement by requirement.
  3. Check out the branch, rerun the acceptance command yourself, and read the exit code.
  4. Weigh each review finding: fix confirmed ones (or send them back), note rejected ones with the reason.
  5. Merge only when the user approved merging for this task.

6. If something fails

Send the implementer one correction packet with the exact failures. Fix small leftovers yourself. After two failed rounds, stop and report what failed and why.

Report to the user: files changed, test results (implementer and your rerun), review findings, and anything you corrected.

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.

Check your setup, free