writing-plans

writing-plans Skill Details

Skill Overview

writing-plans is a Claude skill used to create comprehensive implementation plans before writing code, helping developers break down multi-step tasks into executable, bite-sized steps.

Applicable Scenarios

Projects with multi-step specifications: When you have a complete feature requirement or technical specification document and need to plan a detailed implementation path before starting to code, writing-plans can help you create a structured task breakdown.

Writing guides for team members with no prior experience: When you need to provide detailed implementation guidance for engineers, outsourced developers, or new team members who are unfamiliar with the codebase, this skill ensures that the documentation includes all necessary context and steps.

TDD-driven development workflows: If you want to follow test-driven development (TDD) principles and break each major task into a cycle of “write tests → run tests → minimal implementation → run tests again → commit,” writing-plans is an ideal choice.

Core Features

Bite-Sized Task Breakdown: Breaks complex development tasks into independent steps that can be completed in 2–5 minutes. Each step clearly specifies what to do, how to do it, and how to verify completion. This fine-grained breakdown lets developers see progress quickly and reduces cognitive load.

File Structure Planning: Before defining tasks, it designs clear file boundaries and interfaces. The skill helps determine which files need to be created, which need to be modified, and what responsibility each file should have, ensuring that the code is well organized and easy to maintain.

Complete Task Template Generation: Generates a standardized task structure for each task, including the files involved, step-by-step execution instructions (write tests, run them, implement, verify, commit), complete code examples, and expected output, allowing engineers to execute each task without additional context.

Frequently Asked Questions

How is writing-plans different from ordinary technical documentation?

Traditional technical documentation typically focuses on “what” and “why,” whereas the implementation plans generated by writing-plans focus on “how.” It assumes that the engineer executing the plan is completely unfamiliar with the codebase, so it includes complete code for every step, exact file paths, specific commands, and expected output. Each task is self-contained and can be completed and verified independently.

What should an implementation plan include?

A complete implementation plan must include: a goal statement (a one-sentence description of what will be built), an architectural overview (2–3 sentences explaining the technical approach), the technology stack, and a detailed task list. Each task must clearly identify the files involved (created/modified/tested), provide step-by-step execution instructions (using checkbox syntax), include complete code examples, specify exact run commands and expected output, and provide the commit command. The plan document should be saved in the format docs/superpowers/plans/YYYY-MM-DD-<feature-name>.md.

How do you use a plan generated by writing-plans?

In a Harness that supports subagents, such as Claude Code, use superpowers:subagent-driven-development to execute the plan. Each task is completed by an independent subagent and goes through two levels of review. In environments that do not support subagents, use superpowers:executing-plans to execute the plan in batches within the current session, conducting reviews at key checkpoints. After each task is completed, you should see a clear success signal (passing tests, functional verification, etc.) before proceeding to the next task.

Author

Category

Install