writing-plans

Convert written requirements into a structured, test-driven implementation plan.

9|2|Updated Jan 22, 2026
One-click install
npx skills add https://github.com/DmitriyPolukhin/antigravity-superpower --skill writing-plans-dmitriypolukhin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: writing-plans
Source: https://github.com/DmitriyPolukhin/antigravity-superpower/tree/main/modules/writing-plans
Command: npx skills add https://github.com/DmitriyPolukhin/antigravity-superpower --skill writing-plans-dmitriypolukhin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the problem of turning vague requirements into a concrete, step-by-step engineering plan that an agent can execute safely without touching code prematurely.

Core Features & Use Cases

  • Multi-step plan authoring: Produces a comprehensive implementation plan with explicit file targets, testing steps, and command examples.
  • Assumption-aware engineering guidance: Instructs the implementer to work from zero context and questionable “taste,” ensuring clarity on architecture, files, and verification.
  • TDD-first workflow: Enforces bite-sized tasks that start with failing tests, then implement minimally to pass, and commit frequently.
  • Worktree and execution handoff alignment: Assumes execution happens in a dedicated worktree and includes an explicit handoff to an executing sub-skill.

Quick Start

Tell your agent: “Use the writing-plans skill to generate the implementation plan for my feature spec and save it to docs/plans as a dated markdown file, then ask for subagent-driven execution.”

Frequently Asked Questions about writing-plans

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

FAQPage Schema
How do I convert requirements into a test-driven implementation plan for an AI agent?▼

To convert requirements into a test-driven implementation plan, you provide written specifications to generate a structured plan with explicit file targets, testing steps, and command examples. This prevents premature code changes by enforcing bite-sized tasks that start with failing tests.

How does test-driven development planning prevent premature code changes during project execution?▼

Test-driven development planning prevents premature code changes by enforcing bite-sized tasks where the implementer writes failing tests first, implements minimally to pass, and commits frequently. This ensures behavior verification before any functional code is added.

Can I use git worktrees for executing multi-step software engineering plans?▼

Yes, you can use git worktrees for executing multi-step software engineering plans. The generated plan assumes execution happens in a dedicated worktree-based execution context, ensuring isolation and providing an explicit handoff to a task execution subagent.

What is the best way to structure implementation plans for an agent with limited context?▼

The best way to structure implementation plans for an agent with limited context is to provide explicit file-touch guidance, exact plan headers, and task granularity rules. This ensures deterministic planning and clarity on architecture without relying on the implementer's implicit taste.

Does implementation planning work for multi-step software engineering tasks requiring specific file targets?▼

Yes, implementation planning works for multi-step software engineering tasks requiring specific file targets. It produces a comprehensive plan with explicit file targets, testing steps, and command examples, guiding the implementer safely through zero-context execution.

Why should I use a structured plan instead of letting an agent implement features directly from specs?▼

You should use a structured plan instead of direct implementation to prevent vague requirements from causing premature or incorrect code changes. It applies assumption-aware engineering guidance, ensuring clarity on architecture, files, and verification before any code is touched.