What file is required before running d-plan-fix?

A verified ROOT_CAUSE.md file must exist in the ./.gtd/debug/current/ directory for the planner to generate a fix strategy.

How do I regenerate a plan if one already exists?

You can use the --force flag with the command to overwrite an existing FIX_PLAN.md file.

Why does the skill limit plans to only 2-3 tasks?

This follows the 'Aggressive Atomicity' principle, which forces complex problems to be broken down into small, verifiable steps to reduce risk and prevent side effects.

Does d-plan-fix write the implementation code itself?

No, it focuses on the design and planning phase. It defines interfaces and actions but avoids writing function bodies to maintain high-level architectural oversight.

What is the primary purpose of the d-plan-fix skill?

The skill creates a structured, executable FIX_PLAN.md file based on an existing root cause analysis, ensuring fixes are architectural sound and manageable.

GTD Debug Fix Planner

Name: GTD Debug Fix Planner
Author: Hoang604

byHoang604

0•

生产力与工作流

Generates structured, atomic execution plans to resolve verified software root causes using GTD principles.

The d-plan-fix skill is a specialized architectural tool designed to bridge the gap between root cause analysis and implementation. As part of a 'Get Things Done' (GTD) debugging workflow, it analyzes verified root causes and generates a precise FIX_PLAN.md file. It enforces high-standard engineering principles like aggressive atomicity—limiting plans to a maximum of three tasks—and Gall's Law to prevent over-engineering. This skill is essential for developers who need to ensure that bug fixes address the source of an issue rather than just the symptoms, while maintaining strict side-effect awareness and architectural integrity.

主要功能

01Enforcement of aggressive atomicity (2-3 tasks maximum per plan)

02Automated generation of structured FIX_PLAN.md files

030 GitHub stars

04Defined success criteria and rollback strategies for every fix

05Architectural constraint validation including Single Source of Truth principles

06Built-in side-effect assessment for breaking changes and regressions

使用场景

01Translating a complex root cause analysis into a simple, executable 3-step plan

02Preventing regressions by documenting side effects and verification steps before coding begins

03Standardizing the bug-fixing process to ensure architectural consistency across a project

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hoang604/get-thing-done d-plan-fix

For use in Claude.ai and ChatGPT