writing-plans

Convert requirements into multi-step implementation plans with file mapping and verification.

2|Updated Jun 1, 2017
One-click install
npx skills add https://github.com/carlos-algms/dotfiles --skill writing-plans-carlos-algms
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writing-plans
Source: https://github.com/carlos-algms/dotfiles/tree/main/AI-configs/skills/writing-plans
Command: npx skills add https://github.com/carlos-algms/dotfiles --skill writing-plans-carlos-algms

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you convert a spec into an unambiguous, implementation-ready plan that an agent with no repo context can follow without breaking commit gates or losing track of required constraints.

Core Features & Use Cases

  • Step-by-step implementation planning for agents: Turns multi-step requirements into bite-sized tasks with clear goals, file boundaries, and green-state deliverables.
  • Plan-as-source-of-truth with file-based tracking: Forces the plan to live in a saved file and requires immediate checkbox updates after each verified step.
  • TDD-aware granularity and commit policy alignment: Ensures red/green discipline by keeping failures inside implementation steps, and matches the plan output to the user’s selected commit policy.
  • Verification-first execution: Requires a final verification step and mandates verification-before-completion on the final commands run.

Quick Start

Use the writing-plans skill to generate an implementation plan for a multi-step feature based on a provided spec, without modifying the repo yet.

Frequently Asked Questions about writing-plans

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

FAQPage Schema
How do I create an implementation plan for an AI agent to follow without codebase context?

You create an implementation plan by converting requirements into a multi-step, file-based checklist with explicit file mappings and commit-safe boundaries, ensuring an agent can execute tasks with zero prior repo knowledge.

What is the best way to break down multi-file feature specs into testable agent tasks?

The best way to break down specs is to decompose them into bite-sized, green-state deliverables that maintain TDD red/green discipline, keeping test failures isolated inside implementation steps.

Does writing-plans work with strict commit policies and TDD workflows?

Yes, writing-plans aligns with strict commit policies by matching the plan output to your selected strategy and ensures TDD awareness by keeping failures inside implementation steps before moving to the next task.

How do I track step-by-step progress when an agent executes a complex implementation plan?

You track progress by requiring the plan to live in a saved file and mandating immediate checkbox updates after each verified step, ensuring the plan remains the single source of truth.

Can I use this planning skill for repo-agnostic specs covering multiple subsystems?

Yes, the skill is repo-agnostic and applies directly to specs covering multiple files, subsystems, or testing workflows where tasks must be decomposed into commit-safe, verifiable steps.

What happens if an agent skips the final verification step in the implementation plan?

Skipping verification violates the plan's constraints, which mandates a final verification step and requires verification-before-completion on the final commands run to ensure all deliverables are green.