What happens after a plan is generated?

The skill initiates a review loop for approval and then offers execution options, such as dispatching subagents or performing inline execution using the executing-plans skill.

Does it support custom file structures and commit styles?

Yes, it maps specific file paths, respects existing project patterns, and asks for commit preferences at the start of a session to tailor the plan's granularity.

What is the primary purpose of the Implementation Planner skill?

Its primary purpose is to create a detailed, step-by-step technical roadmap before any code is touched, ensuring that developers or AI agents have a clear, test-driven path to follow.

How does this skill handle Test-Driven Development (TDD)?

It breaks every task into specific TDD cycles: writing a failing test, verifying the failure, implementing minimal code, and verifying the pass, ensuring high code quality.

Where are the implementation plans saved?

Plans are typically saved to the docs/superpowers/plans/ directory as Markdown files, though user-defined location preferences can override this default.

Implementation Planner

Name: Implementation Planner
Author: nocksock

bynocksock

•

開発者ツール

Generates granular, TDD-focused implementation plans and architectural roadmaps before executing code changes.

Implementation Planner is a high-precision skill designed to bridge the gap between abstract requirements and production-grade code. It decomposes complex specs into bite-sized, 2-5 minute tasks centered around Test-Driven Development (TDD) and the DRY/YAGNI principles. By mapping out file structures, defining clear interfaces, and enforcing a mandatory review loop, this skill ensures that every modification is intentional, testable, and architecturally sound. It is particularly effective for hand-offs between planning and execution phases, whether using subagents or inline workflows.

主な機能

01Granular 2-5 minute task decomposition for maximum execution reliability

02Test-Driven Development (TDD) task mapping with failing/passing test steps

03Automated architectural scope checks and subsystem decomposition

04Integrated plan review loop with dedicated reviewer subagents

051 GitHub stars

06Session-wide commit preference management and execution handoff options

ユースケース

01Decomposing monolithic requirements into manageable, atomic implementation steps

02Planning high-stakes refactors that require strict test coverage

03Scaffolding complex new features in large, unfamiliar codebases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nocksock/superpowers writing-plans

For use in Claude.ai and ChatGPT