writing-skills

Design and validate skill documentation with frontmatter and pressure-testing workflows.

Updated Apr 3, 2026

Search Tags:#tdd #documentation #prompt design #skill authoring #red-green-refactor #subagent testing

Ready to run

npx skills add https://github.com/norrbacka/threejs_codex_base --skill writing-skills-norrbacka

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: writing-skills
Source: https://github.com/norrbacka/threejs_codex_base/tree/main/.codex/superpowers/skills/writing-skills
Command: npx skills add https://github.com/norrbacka/threejs_codex_base --skill writing-skills-norrbacka

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps authors create, edit, and verify other skills so they work reliably before deployment, reducing vague instructions, missed edge cases, and agent drift.

Core Features & Use Cases

Skill authoring guidance: Shape clear names, trigger-focused descriptions, and concise operational bodies.
Verification workflow: Test skills with realistic pressure scenarios, then refine them based on observed failures.
Use case: When a new skill keeps getting ignored, this Skill helps diagnose discovery problems, tighten the instructions, and validate that agents comply under pressure.

Quick Start

Tell the AI to use the writing-skills skill when creating, editing, or testing a new skill before deployment.

Frequently Asked Questions about writing-skills

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

FAQPage Schema

How do I validate skill documentation for reliable agent behavior before deployment?▼

You validate skill documentation by applying baseline-plus-pressure verification with subagents to test instructions against realistic scenarios. This process designs and verifies skill triggers, operational guidance, and frontmatter to prevent agent drift and missed edge cases before deployment.

Why does my AI agent keep ignoring my custom skill instructions?▼

Agent drift and ignored instructions often stem from vague skill descriptions or missing trigger-focused guidance. Diagnosing discovery problems requires tightening operational bodies and pressure-testing skill instructions with subagents to ensure agents comply under realistic deployment conditions.

What is the best way to structure skill authoring for consistent agent execution?▼

The best way to structure skill authoring is by ensuring clear frontmatter, concise operational guidance, and trigger-focused descriptions. This structure shapes reliable agent behavior by reducing ambiguity and providing precise instructions for the agent to follow during execution.

Can I use TDD principles like red-green-refactor to test AI prompt design?▼

Yes, you can apply red-green-refactor testing principles to prompt design by using subagent testing workflows. This approach pressure-tests skill instructions with baseline scenarios, observes failures, and refines the documentation iteratively until the agent behavior meets deployment standards.

How do I pressure-test skill instructions to catch edge cases before deployment?▼

You pressure-test skill instructions by running subagents against realistic pressure scenarios and refining the skill based on observed failures. This verification workflow diagnoses discovery problems, tightens vague instructions, and validates that agents handle edge cases reliably before deployment.

Do I need specific dependencies to start editing and validating existing skills?▼

No specific dependencies are required to start editing and validating existing skills. The process relies on clear frontmatter, concise operational guidance, and baseline-plus-pressure verification with subagents to shape trigger-focused descriptions and ensure reliable agent behavior before deployment.