writing-plans

Create implementation plans with TDD tasks, file paths, and .tasks.json.

Updated Feb 23, 2026
One-click install
npx skills add https://github.com/rishikanthc/ml-superpowers --skill writing-plans-rishikanthc
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writing-plans
Source: https://github.com/rishikanthc/ml-superpowers/tree/main/skills/writing-plans
Command: npx skills add https://github.com/rishikanthc/ml-superpowers --skill writing-plans-rishikanthc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Converts vague specifications or requirements into a concrete, executable implementation plan so engineers can start coding without ambiguity, missing steps, or poor test coverage.

Core Features & Use Cases

  • Structured, bite-sized tasks: Break features into 2–5 minute actionable steps: failing test, run, implement minimal code, run, commit.
  • Worktree and TDD-first workflow: Assumes a dedicated worktree and enforces TDD, exact file paths, commit guidance, and frequent checkpoints.
  • Native task integration: Creates and links native tasks with blockedBy relationships and produces a co-located .tasks.json to resume execution across sessions.
  • Use Case: Take a product spec and produce a saved docs/plans/YYYY-MM-DD-feature.md with matching native tasks and clear execution handoff.

Quick Start

Create an implementation plan from a spec, save it to docs/plans/YYYY-MM-DD-<feature>.md, and generate matching native tasks with a co-located .tasks.json.

Frequently Asked Questions about writing-plans

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn a feature specification into a TDD implementation plan?

To turn a feature specification into a TDD implementation plan, provide your requirements to generate a structured, bite-sized task list. The output includes failing tests, exact file paths, run commands, and commit steps saved as a dated markdown file with matching native tasks.

What is the best way to break multi-step engineering work into bite-sized tasks?

The best way to break multi-step engineering work into bite-sized tasks is using a TDD-first workflow that segments features into 2–5 minute actionable steps. Each step cycles through writing a failing test, running it, implementing minimal code, and committing.

How does native task tracking integrate with implementation plans?

Native task tracking integrates with implementation plans by creating linked tasks with blockedBy relationships and a co-located .tasks.json file. This setup allows you to resume execution seamlessly across sessions while maintaining dependency awareness.

Do I need a dedicated worktree to use this implementation planning workflow?

Yes, a dedicated worktree is required for this implementation planning workflow. The process assumes a worktree-first approach to isolate changes, enforce test-driven development checkpoints, and maintain exact commit guidance throughout the feature build.

Can I save implementation plans as markdown files in my repository?

Yes, you can save implementation plans as markdown files in your repository. The process automatically generates a saved docs/plans/YYYY-MM-DD-feature.md file containing the full structured task breakdown, file paths, and execution handoff details.

What limitations should I expect when generating TDD-style tasks from specifications?

A limitation when generating TDD-style tasks is that the execution handoff is constrained to an AskUserQuestion mechanism for choice selection. Additionally, the workflow strictly enforces frequent checkpoints and minimal code implementation, which may not suit non-test-driven development environments.