name: skill-creator
description: Create new agent skills with proper structure, frontmatter, and progressive disclosure. Use when building a skill from scratch, initializing skill directories, validating skill structure, or packaging skills for distribution.

Skill Creator

Create well-structured agent skills.

When This Triggers

• User wants to create a new skill
• User asks about skill structure or SKILL.md format
• User needs to validate or package an existing skill
• User is converting a workflow into a skill

Quick Start

Initialize a New Skill

python scripts/init_skill.py <skill-name> --path <output-directory>

Creates:

skill-name/
├── SKILL.md
├── scripts/
├── references/
└── assets/

Validate a Skill

python scripts/quick_validate.py <path/to/skill-folder>

Checks frontmatter, naming conventions, structure.

Package for Distribution

python scripts/package_skill.py <path/to/skill-folder> [output-directory]

Creates a .skill zip archive.

Skill Structure

skill-name/
├── SKILL.md              # Required: instructions and metadata
├── scripts/              # Optional: executable code
├── references/           # Optional: docs loaded as needed
└── assets/               # Optional: files used in output

Delete directories you don't need.

Writing SKILL.md

Frontmatter (Required)

---
name: skill-name
description: What it does and WHEN to use it
---

• name: hyphen-case, lowercase, max 64 chars
• description: max 1024 chars, no angle brackets

The description is how the agent decides to trigger your skill. Include:
- What the skill does
- Specific scenarios that should activate it
- File types or task types it handles

Body

Keep under 500 lines. Structure:
1. Overview (1-2 sentences)
2. When this triggers
3. Quick start / main workflow
4. Reference links for deep dives

Progressive Disclosure

Only load what's needed:

1. Metadata — Always visible (~100 words)
2. SKILL.md body — When triggered (<5k words)
3. References — On demand (unlimited)
4. Scripts — Execute without loading

When to Use Each

Creation Process

1. Understand with examples — What should this skill support?
2. Plan reusable contents — What scripts, references, assets help?
3. Initialize — Run init_skill.py
4. Implement — Write SKILL.md, add resources
5. Validate — Run quick_validate.py
6. Iterate — Use it, notice struggles, improve

Common Patterns

Sequential Workflow

1. Analyze input (run analyze.py)
2. Transform (run transform.py)
3. Validate output (run validate.py)

Conditional Workflow

Determine the task type:
- **Creating new?** → Follow "Creation" section
- **Editing existing?** → Follow "Editing" section

Reference Navigation

## Advanced features
- **Forms**: See references/forms.md
- **API**: See references/api.md

Scripts

This skill includes helper scripts in scripts/:

• init_skill.py — Initialize new skill from template
• quick_validate.py — Validate skill structure
• package_skill.py — Package for distribution

See references/skill-anatomy.md for detailed specs.