Back to Skills

planning

Planning work in small, known-good increments. Use when starting significant work or breaking down complex tasks.

524stars71forksUpdated 2/18/2026

Security Assessment

Safe(100/100)
Security Score100/100

About planning

The 'planning' skill is designed to help teams break down complex tasks and start significant work in small, manageable increments. The core principle is that all work must be done in small, known-good steps, ensuring that each increment leaves the codebase in a working state with all tests passing. This approach minimizes risks by ensuring that each step is independently deployable and clearly defined. It emphasizes maintainability, clarity, and test-driven development (TDD) practices, making it easier for teams to track progress and adjust as needed.

Key features of this skill include the use of a three-document model to maintain planning documentation: PLAN.md for outlining goals, WIP.md for tracking progress, and LEARNINGS.md for capturing discoveries. These documents work in conjunction, guiding the project through structured steps while also documenting key insights and blockers along the way. The 'known-good increment' methodology ensures that each step is small enough to be described succinctly, independently deployable, and aligned with testing and refactoring principles.

This skill is useful for developers and teams working on significant or complex tasks where clear tracking, approval processes, and incremental progress are essential. It's particularly beneficial for those working in agile environments or following TDD principles, as it promotes clear task breakdown, accountability, and consistent quality checks throughout the development process.

FAQ

How should I break down large tasks into smaller increments?

Each step must leave the codebase in a working state, with all tests passing. The task should be described in one sentence and should not involve more than three files. If a step can't be described succinctly, break it down further.

What is the purpose of the three-document model?

The three-document model helps organize and track significant work. PLAN.md outlines the goals and acceptance criteria, WIP.md tracks current progress, and LEARNINGS.md documents discoveries or challenges faced during development.

What are the requirements for committing work?

You must get user approval before committing. After completing a step, verify that all tests and static analysis pass, update the WIP.md, and capture any insights in the LEARNINGS.md. Only commit after receiving explicit approval.

How does TDD integrate with this planning skill?

Each step follows the RED-GREEN-REFACTOR cycle: write a failing test first (RED), write the minimum code to pass (GREEN), and then refactor if necessary (REFACTOR), ensuring all tests pass before proceeding.

What happens to the LEARNINGS.md document after the feature is completed?

The LEARNINGS.md is merged into relevant documents like CLAUDE.md for gotchas and patterns or ADRs for architectural decisions and is then deleted after the feature is finished.

Install planning

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill