write-a-skill

Create Claude Code agent skills with SKILL.md and progressive disclosure.

6|1|Updated Apr 5, 2026
One-click install
npx skills add https://github.com/snewhouse/aa-ma-forge --skill write-a-skill-snewhouse
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: write-a-skill
Source: https://github.com/snewhouse/aa-ma-forge/tree/main/claude-code/skills/write-a-skill
Command: npx skills add https://github.com/snewhouse/aa-ma-forge --skill write-a-skill-snewhouse

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps users create new agent skills with consistent structure and clear, discoverable instructions so others (and the agent) can reliably understand when to use them.

Core Features & Use Cases

  • Requirement gathering: Prompts the user to clarify domain, use cases, and whether executable scripts or only instructions are needed.
  • Skill drafting workflow: Produces a root SKILL.md with progressive disclosure, plus optional additional reference files when content grows large.
  • Operational structure guidance: Recommends a standard directory layout (SKILL.md plus optional reference/examples/scripts) to keep skills maintainable and reusable.
  • Use Case: Writing a new Claude Code skill for a recurring workflow like generating consistent project summaries, where you want the agent to load the right skill from the description and then follow the documented processes.

Quick Start

Use the write-a-skill skill to draft a new skill named "my-skill" with a clear SKILL.md description and workflow sections for the domain "X".

Frequently Asked Questions about write-a-skill

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

FAQPage Schema
How do I create a new Claude Code skill from scratch?

To create a Claude Code skill, draft a root SKILL.md file with YAML frontmatter containing the name and description, then add workflow sections and optional reference files for progressive disclosure.

What is progressive disclosure in agent skill creation?

Progressive disclosure in skill creation splits instructions across a root SKILL.md and optional reference files, allowing the agent to load detailed documentation on demand rather than all at once.

Do I need to include deterministic scripts in my SKILL.md structure?

Including deterministic scripts in your SKILL.md structure is optional, decided during requirement gathering based on whether the skill needs executable automation or only documented instructions.

Can I bundle reference files and examples with a Claude Code skill?

You can bundle reference files and examples with a Claude Code skill by using a conventional directory layout alongside the root SKILL.md to keep complex content maintainable and reusable.

What is the best way to structure YML frontmatter for an agent skill?

The best way to structure YML frontmatter for an agent skill is to include required name and description fields at the top of SKILL.md, ensuring the agent can discover and load the right skill.

When should I split skill instructions into separate reference files?

You should split skill instructions into separate reference files when the content grows too large for a single document, applying progressive disclosure to maintain readability and on-demand loading.