skill-writer

Evaluate and refine agent skill text against structural rubrics.

Updated Apr 30, 2026

Search Tags:#workflow #documentation #skills #agent #evaluation #rubric

Ready to run

npx skills add https://github.com/kazukitcy/skills --skill skill-writer-kazukitcy

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-writer
Source: https://github.com/kazukitcy/skills/tree/main/skill-writer
Command: npx skills add https://github.com/kazukitcy/skills --skill skill-writer-kazukitcy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill addresses the challenge of creating agent skills that are predictable, robust, and maintainable, preventing common issues like premature completion, vague instructions, and poor information hierarchy.

Core Features & Use Cases

Rubric-Based Evaluation: Assesses skills against a comprehensive rubric covering invocation, rule form, and information hierarchy.
Checklist-Driven Review: Provides a structured, step-by-step review process to identify blocking issues before distribution.
Use Case: Use this skill when drafting a new skill from scratch or auditing an existing skill to ensure it adheres to professional standards and avoids common failure modes.

Quick Start

Use the skill-writer to evaluate the current skill directory against the rubric and provide a report of findings.

Frequently Asked Questions about skill-writer

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

FAQPage Schema

How do I evaluate agent skills against a structural rubric?▼

To evaluate agent skills against a structural rubric, assess documentation for invocation fit, rule form, and information hierarchy to ensure predictability and operational clarity. This structured review identifies blocking issues and prevents common failure modes like vague instructions.

What is the best way to refine agent workflows and prevent premature completion?▼

Refining agent workflows to prevent premature completion requires applying a checklist-driven review to skill text and documentation. This process validates structural rubrics, enforces operational clarity, and ensures instructions are predictable, robust, and maintainable before distribution.

How do I audit existing agent documentation for common failure modes?▼

Auditing existing agent documentation for common failure modes involves evaluating the text against a comprehensive structural rubric. This checklist-driven review identifies vague instructions, poor information hierarchy, and blocking issues to ensure adherence to professional standards.

Can I use a checklist-driven review to draft new agent skills from scratch?▼

Yes, you can use a checklist-driven review to draft new agent skills from scratch by validating the documentation as it is created. This ensures invocation fit, proper rule form, and correct information hierarchy are established before the skill is distributed.

Why does my agent skill have poor information hierarchy and vague instructions?▼

Agent skills suffer from poor information hierarchy and vague instructions when they are not evaluated against a comprehensive structural rubric during the drafting phase. Applying a checklist-driven validation process ensures operational clarity and prevents these common failure modes.

When do I need a rubric-based evaluation for agent skill lifecycle management?▼

You need a rubric-based evaluation for agent skill lifecycle management whenever you are drafting, reviewing, or validating skill documentation. This ensures the skills maintain predictability, operational clarity, and adherence to structural standards throughout their distribution lifecycle.