skill-creator

Skill Creator

This skill guides you through creating effective, reusable AI agent skills.

Core Principles

1. Concise is Key: Only add context the agent doesn't already have. Challenge each piece: "Does this justify its token cost?"
2. Appropriate Freedom: Match specificity to task fragility - use guardrails for fragile operations, flexibility for adaptable tasks.
3. Progressive Disclosure: Keep SKILL.md lean (<500 lines), split detailed content into references/.

Skill Structure

skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name, description)
│   └── Markdown instructions
└── Optional Resources
    ├── scripts/      - Executable code
    ├── references/   - Documentation for context
    └── assets/       - Files for output (templates, etc.)

Creation Workflow

Step 1: Analyze the Task Solution

Before creating, identify:

• Core workflow: What steps solved the problem?
• Reusable patterns: What would help in similar tasks?
• Trigger conditions: When should this skill activate?

Step 2: Initialize the Skill

Run the init command to create the template:

npx skills init <skill-name>

This creates a skill directory with SKILL.md template and resource directories.

Step 3: Write the SKILL.md

Frontmatter (Required)

---
name: skill-name
description: What the skill does AND when to use it. Include specific triggers.
---

Critical: The description is the primary trigger mechanism. Include:

• What the skill does
• Specific scenarios that trigger it
• File types or task patterns it handles

Body Structure

Choose a structure based on skill type:

Workflow-Based (sequential processes):

# Skill Title
## Overview
## Workflow Decision Tree
## Step 1: [Action]
## Step 2: [Action]

Task-Based (tool collections):

# Skill Title
## Overview
## Quick Start
## Task: [Operation 1]
## Task: [Operation 2]

Step 4: Add Resources (If Needed)