agentleFS
sxntixgosxntixgo/frugal-skills/skills/writing-plans/SKILL.md

writing-plans

Skill0 starsChanged 4 months agoMITView on GitHub

Turns a spec or design into a phased implementation plan with verifiable steps and review gates, before any code is written. Use when you have requirements or a design for a multi-step task and need a plan, or before starting implementation.

188 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use chinese-code-review, chinese-commit-conventions 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.
  • Leodorareluctant259Skill

    chinese-code-review

    中文代码审查规范——在保持专业严谨的同时,用符合国内团队文化的方式给出有效反馈

    Also in 8 projects that use this

  • Leodorareluctant259Skill

    chinese-commit-conventions

    中文 Git 提交规范 — 适配国内团队的 commit message 规范和 changelog 自动化

    Also in 8 projects that use this

  • ryanportfolioSkill

    addskill

    Create, import, update, or install repo or personal skills: runtime ownership, resources, discovery checks.

    Also in 6 projects that use this

  • ryanportfolioSkill

    apply-firmware

    Bring current folder (plain folder, existing repo, old or partial firmware copy) up to full Harness Firmware in place. Mirror external repo → adopt-repo.

    Also in 6 projects that use this

  • +2 more from Leodorareluctant259
  • +6 more from ryanportfolio

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 170 edited versions we read keep these.

  • 5

    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:

    Read the spec/design and the relevant code, then produce a phased plan where each phase is independently verifiable and ordered by …

    line 13
  2. Very specific:

    Write the implementation code → that's software-development.

    line 18
  3. Distinctive:

    Read the spec/design AND the relevant existing code.

    line 28
  4. Distinctive:

    Prefer the smallest slice that delivers real behavior first.

    line 35
  5. Very specific:

    Write the finished plan to docs/plans/YYYY-MM-DD-<topic>-plan.md.

    line 47
  6. Distinctive:

    Every phase is independently verifiable and ordered by dependency.

    line 62
  7. Distinctive:

    Every task states what, where, and how to verify.

    line 63

“Common” is measured: kept by at least 6 in 10 of the 170 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

Writing Plans

Produce a plan someone could execute without you. Every task says what to change, where, and how to verify it. Plan on paper first — don't start coding.

Scope

This skill DOES:

  • Read the spec/design and the relevant code, then produce a phased plan where each phase is independently verifiable and ordered by dependency.
  • Give each task a concrete verification step and mark review/STOP gates.

This skill does NOT:

  • Write the implementation code → that's software-development.
  • Invent requirements from a vague ask → if the spec is unclear, stop and clarify the requirements first, before planning.
  • Pad the plan with speculative future work → YAGNI; list it as out-of-scope.

Steps

Surface these as TodoWrite items. The plan file itself is the checkpoint — write it incrementally to docs/plans/YYYY-MM-DD-<topic>-plan.md so work resumes.

1. Ground the plan ⛔ STOP if the spec is unclear

Read the spec/design AND the relevant existing code. Restate the goal and the observable success criteria in a sentence or two. If the ask is ambiguous or under-specified, STOP and clarify — do not invent requirements.

2. Decompose into phases

Break the work into small phases, each independently verifiable and, ideally, independently shippable. Order them by dependency so each builds on the last. Prefer the smallest slice that delivers real behavior first.

3. Detail each task

For every task write: what to change, where (file/seam), and how to verify it — a test to write/run, or a behavior to observe. A task with no verification step is not a real task. See references/plan-structure.md.

4. Mark gates, risks, and out-of-scope

Insert review/STOP checkpoints between phases. List risks and assumptions, an explicit Out of scope (YAGNI) section, and any open questions.

5. Write the plan and hand off

Write the finished plan to docs/plans/YYYY-MM-DD-<topic>-plan.md. Point the user to software-development to execute it phase by phase.

Red flags (stop and correct)

If you notice…Do this
A task with no verification ("improve X", "handle errors")Make it concrete + add how to verify
A giant phaseSplit into smaller independently verifiable slices
Planning without reading the codeRead the actual code first (Step 1)
Speculative "future-proofing" in the planMove it to Out of scope (YAGNI)
Phases in arbitrary orderOrder by dependency; smallest real slice first
Planning from a vague specSTOP and clarify the requirements before planning
Starting to write implementation codeStop — this skill plans; software-development builds

Done when

  • Every phase is independently verifiable and ordered by dependency.
  • Every task states what, where, and how to verify.
  • Risks, assumptions, out-of-scope, and open questions are listed.
  • The plan is written to docs/plans/.

References (load only when needed)

  • references/plan-structure.md — anatomy of a good plan: phase/task format, verification, gates, and out-of-scope.

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.