skill-refine

Extract actionable principles from user feedback and update SKILL.md files.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/kentoku24/project-guild --skill skill-refine
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-refine
Source: https://github.com/kentoku24/project-guild/tree/main/docs/skills/skill-refine
Command: npx skills add https://github.com/kentoku24/project-guild --skill skill-refine

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The Skill Refine process reduces drift by converting user feedback into abstract, reusable principles and updates SKILL.md and related files accordingly, avoiding indiscriminate rule expansion.

Core Features & Use Cases

  • Extracts core, domain-agnostic principles from concrete feedback to guide future SKILL updates.
  • Updates YAML frontmatter (name, description) and core sections to reflect the extracted abstraction, keeping the skill structure DRY.
  • Applies to any skill documented with SKILL.md when users provide feedback or when the /skill-refine command is invoked.

Quick Start

Use /skill-refine with the target skill name and concrete feedback to update SKILL.md and related files.

Frequently Asked Questions about skill-refine

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

FAQPage Schema
How do I refine SKILL.md files using user feedback without adding unnecessary rules?

To refine SKILL.md files, you extract abstract, domain-agnostic principles from concrete feedback. This updates the skill's core guidance and YAML frontmatter while preserving abstraction and preventing rule bloat.

What is the best way to update YAML frontmatter name and description from feedback?

Updating YAML frontmatter name and description involves extracting actionable principles from user feedback. The refine process applies these abstractions to keep the skill structure DRY and accurately reflect the updated guidance.

How do I extract actionable principles from concrete feedback to update skill documentation?

Extracting actionable principles from concrete feedback requires distilling the input into domain-agnostic guidance. These extracted principles then drive updates to SKILL.md and related files, reducing drift and avoiding indiscriminate rule expansion.

Can I apply skill refinement to any skill documented with SKILL.md?

Yes, you can apply skill refinement to any skill documented with SKILL.md. By invoking the /skill-refine command with the target skill name and concrete feedback, the process updates the core sections and frontmatter accordingly.

Why does skill documentation drift and how do I prevent it?

Skill documentation drifts when feedback is added as indiscriminate rules rather than abstracted principles. You prevent drift by converting concrete feedback into reusable, DRY principles that update SKILL.md core sections while preserving abstraction.