ワンクリックで
create-implementation-plan
Instructions for writing good implementation plans artifacts. Use when specifically requested.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Instructions for writing good implementation plans artifacts. Use when specifically requested.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Publishes a clean normal feature or task branch and creates its pull request on explicit human request.
Instructions for writing good design plan artifacts. Use when specifically requested.
Instructions for executing implementation plans. Use when specifically requested.
Instructions for executing a task from an implementation plans. Use when specifically requested.
Instructions for investigating a codebase to answer a specific question.
Reviews code changes directly against quality standards. Use when no implementation plan or journal exists.
| name | create-implementation-plan |
| description | Instructions for writing good implementation plans artifacts. Use when specifically requested. |
Read the provided design plan artifact and use it to produce an implementation plan artifact.
Use this skill to translate an approved design plan into a step-by-step implementation plan.
It is always the second artifact stage in the run-feature workflow, and is also used
in run-bug-fix (where a bug report takes the place of a design plan).
This skill is a sub-skill called by orchestrators:
run-feature — stage 2 (planning)run-bug-fix — planning stagerun-fix — fix plan stageDo not use this skill to write design plans or code. Do not confuse with create-design-plan,
which produces the higher-level artifact that this skill consumes.
Your prompt must include:
If not provided, ask before proceeding. Do not guess.
Implementation plans are lower-level artifacts that expand higher-level design plans into a more detailed, step-by-step plan to execute the implementation.
Implementation plans describe the process to implement a feature, a refactor, a new project, etc. They are specific, step-by-step instructions that produce complete, testable, and documented implementations. They should describe specific code changes in particular files, modules, functions, etc. They should itemize tests to build at multiple levels: unit, integration, end-to-end, etc.
The plan should break the process into succinct, concrete steps. The outcome of each step should be scoped and verifiable.
Plans should be written assuming the implementor has little context for the project. You should assume the implementor is a good engineer, but needs careful guidance in project standards, style, structure, and taste. Do not assume domain expertise.
The plan artifact must operate within the project standards and structure. If the plan violates either, reviewers will call it out and you will look bad. Have care.
The plan MUST conform to the structure defined in the template! Plans that deviate will be flagged by the plan reviewer with Critical Findings. Avoid that embarrassment by making sure your plan is conformant.
The work is complete after the implementation plan is created. The work does not include writing any code.
Write the implementation plan to {project-dir}/implementation-plan.md, where {project-dir} is the
directory containing the design plan. Read .agents/artifacts/implementation-plan/description.md for the
canonical section definitions, and render .agents/artifacts/implementation-plan/template.md.j2 to produce
the initial file. Replace all dummy content — every line drawn from the retro encabulator — with real
content for this project. The rendered file must contain no placeholder text when submitted. Supply the
path to your caller on completion.
Before writing, read and follow ~/.agents/instructions/markdown.md. All heading spacing,
list formatting, line length, and code fence rules apply to every plan artifact.