agentleFS
ac5tinac5tin/skills/skills/implementation-plan/SKILL.md

implementation-plan

Skill0 starsChanged 35 days agoMITView on GitHub

Use when writing, revising or checking an implementation plan for a delegated task. Defines the required structure and the definition-of-done mapping.

24 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use architecture-review and security-review.

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.
  • SentrySkill

    architecture-review

    Staff-level codebase health review. Finds monolithic modules, silent failures, type safety gaps, test coverage holes, and LLM-friendliness issues.

    Also in 3 projects that use this

  • SentrySkill

    security-review

    Finds exploitable application security vulnerabilities in code changes. Use for Warden security scans, appsec review, OWASP-style checks, authentication or authorization bugs, injection, XSS, SSRF, path traversal, secrets, unsafe crypto, webhook verification, open redirects, or sensitive data exposure.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 22 edited versions we read keep these.

  • 6

    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:

    Every plan follows this structure before the leader approves it.

    line 8
  2. Distinctive:

    Follow-ups and open questions — parked work only.

    line 33
  3. Distinctive:

    Every goal has at least one done criterion.

    line 41
  4. Distinctive:

    Every done criterion names one check.

    line 42
  5. Distinctive:

    Every check maps back to a goal.

    line 43
  6. Distinctive:

    No silent blockers, no TBD owners, no work parked on a peer that the driver can do.

    line 53

“Common” is measured: kept by at least 6 in 10 of the 22 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

Implementation plan

Every plan follows this structure before the leader approves it. The plan describes the current design only. Keep one living artifact and revise it in place.

Generic by design: no company names, accounts, storage products, people or role names. The template supplies where the plan lives and who approves it.

Front matter

Title, version, date in the leader's timezone, author, status, task reference, repositories.

Sections, in order

  1. Background — why this exists, with links to the source material.
  2. Goals and measurable outcomes — separate must-have from nice-to-have.
  3. Terminology — terms unique to this task.
  4. Scope — in scope, and explicit non-goals.
  5. Architecture and approach — at least one high-level diagram, inline with alt text.
  6. Detailed design — contracts, control gates, data path, external services, errors and observability.
  7. Data and storage changes — migrations, apply path, rollback; say "none" when none.
  8. Affected systems — table: system, change type, notes.
  9. Autonomy and hard gates — ownership statement, equip table, owner handoff packs.
  10. Implementation steps — ordered units the driver executes.
  11. Verification — pass/fail bars, one check per done criterion.
  12. Rollout and rollback — environment order, enablement, triggers, procedure, authority, recovery check.
  13. Trade-offs and decisions — decision, alternatives, reason, consequence.
  14. Limitations — honest limits of the chosen design.
  15. Risks — likelihood, impact, mitigation, owner.
  16. Follow-ups and open questions — parked work only. No development-stopping item here.
  17. Definition of done and responsibilities — one row per goal with its criterion and check.
  18. Appendix — changelog only, one line per version.

Definition of done

The mapping is the gate:

  • Every goal has at least one done criterion.
  • Every done criterion names one check.
  • Every check maps back to a goal. Remove or map an orphan check.
  • Each check states the exact command or observation, expected result, pass/fail bar, inputs, starting state, and evidence location.
  • The driver can run every check without human action.
  • A goal without a check is not done.

Ownership

  • The driver is the default operator: code, scripts, migrations, rollout, verification.
  • Peers approve bars and advise. They do not become the operator.
  • A human step appears only when no bot has access. It carries best time, exact commands, inputs, expected output, done signal and rollback.
  • No silent blockers, no TBD owners, no work parked on a peer that the driver can do.

Quality

  • One living plan, revised in place. No new artifact per version.
  • The title is stable. The version lives in the front matter, not the title.
  • Prune superseded drafts, duplicate uploads and obsolete figures.
  • Front matter version and status are current.
  • Rejected alternatives stay short, in the decisions section.
  • Figures have alt text or a caption, and sit in the same place as the plan.
  • Cite the source and date for any external claim.

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.