Skill Creator Enhanced

Create production-grade skills that extend Claude's capabilities.

What This Skill Does

• Guides creation of new skills from scratch
• Helps improve existing skills to production quality
• Provides patterns for all skill types (Builder, Guide, Automation)
• Ensures skills meet validation criteria

What This Skill Does NOT Do

• Create the actual domain content (user provides that)
• Test skills in production environments
• Deploy or distribute skills
• Handle skill versioning/updates after creation

---

Required Clarifications

Before creating a skill, ask:

1. Skill Type

"What type of skill are you creating?"

• Builder → Creates artifacts (widgets, code, documents)
• Guide → Provides instructions (how-to, tutorials)
• Automation → Executes workflows (file processing, deployments)

2. Domain

"What domain or technology does this skill cover?"

• Example: "ChatGPT widgets", "PDF processing", "BigQuery analytics"

3. Concrete Examples

"Can you give 2-3 examples of how this skill would be used?"

• Example prompts users would give
• Expected outputs for each

4. Existing Resources

"Do you have any existing scripts, templates, or documentation to include?"

• Scripts for scripts/
• Templates for assets/
• Reference docs for references/

Optional Clarifications

5. Official Docs: "Are there official documentation URLs for this domain?"
6. Standards: "Are there industry standards to enforce?" (WCAG, OWASP, etc.)

---

Core Principles

Concise is Key

Context window is a public good. Challenge each piece of information:

• "Does Claude really need this explanation?"
• "Does this paragraph justify its token cost?"

Prefer concise examples over verbose explanations.

Appropriate Freedom

Match specificity to task fragility:

Freedom Level	When to Use	Example
High	Multiple approaches valid	"Choose your preferred style"
Medium	Preferred pattern exists	Pseudocode with parameters
Low	Operations are fragile	Exact scripts, few parameters

Progressive Disclosure

Three-level loading system:

1. Metadata (~100 words) - Always in context
2. SKILL.md body (<500 lines) - When skill triggers
3. References (unlimited) - As needed by Claude

---

Anatomy of a Skill

skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name, description)
│   └── Markdown instructions
└── Bundled Resources (optional)
    ├── scripts/      - Executable code
    ├── references/   - Documentation loaded as needed
    └── assets/       - Templates, images, fonts

SKILL.md Requirements

Component	Requirement
Line count	<500 lines (extract to references/)
Frontmatter	name + description with triggers
Form	Imperative ("Do X" not "You should X")
Scope	What it does AND does not do

What NOT to Include

• README.md (SKILL.md IS the readme)
• CHANGELOG.md
• LICENSE (inherited from repo)
• Duplicate information

---

Skill Creation Process

Step 1: Understand with Examples

Ask clarifying questions (see Required Clarifications above).

Conclude when you have:

• Clear skill type (Builder/Guide/Automation)
• 2-3 concrete usage examples
• Known resources to include

Step 2: Plan Resources

For each example, identify:

Resource Type	When to Create
scripts/	Same code rewritten repeatedly, deterministic reliability needed
references/	Documentation Claude should reference while working
assets/	Templates, images, boilerplate for output

Step 3: Initialize

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

Creates template with proper structure.

Step 4: Implement

1. Start with resources - Create scripts/, references/, assets/ files
2. Test scripts - Run to verify they work
3. Write SKILL.md - Follow patterns below

SKILL.md Patterns

Frontmatter:

---
name: skill-name
description: What it does + when to use + triggers. Use when users ask to...
---

Body structure:

1. What this skill does / does not do
2. Required clarifications (for Builder skills)
3. Core workflow steps
4. Reference files table
5. Output checklist

See references/skill-patterns.md for complete patterns.

Step 5: Add Quality Elements

For production-grade skills, include:

Element	Where	Purpose
Clarification questions	SKILL.md	Prevent wrong assumptions
Official doc links	references/	Enable latest patterns
Enforcement checklists	SKILL.md or references/	Ensure compliance
Output checklist	SKILL.md	Quality gate before delivery
Error handling	references/	Graceful failures

See references/quality-patterns.md for implementation details.

Step 6: Package

scripts/package_skill.py <path/to/skill-folder>

Validates and creates .skill file.

Step 7: Iterate

Use skill on real tasks → Notice issues → Update and retest.

---

Output Checklist

Before delivering a skill, verify:

Structure

• SKILL.md exists and <500 lines
• Frontmatter has name + description with triggers
• No README.md, CHANGELOG.md in skill directory
• Progressive disclosure (details in references/)

Content

• Imperative form throughout
• Scope clarity (does / does not do)
• Clear workflow steps
• No verbose explanations

User Interaction (Builder skills)

• Required clarifications section
• Optional clarifications separated
• Context awareness guidance

Documentation

• Official doc links (if applicable)
• Reference files for complex details
• "When to read" table for references

Domain Standards (if applicable)

• Must Follow checklist
• Must Avoid list
• Output quality gate

Technical (if applicable)

• Error handling guidance
• Dependencies documented
• Scripts tested

---

Reference Files

File	When to Read
references/skill-patterns.md	SKILL.md structure and examples
references/quality-patterns.md	Clarifications, enforcement, checklists
references/technical-patterns.md	Error handling, security, dependencies
references/workflows.md	Sequential and conditional workflows
references/output-patterns.md	Template and example patterns

---

Quick Reference

Must Follow

• SKILL.md <500 lines
• Frontmatter with triggers
• Imperative form
• No extraneous files
• Progressive disclosure

Must Avoid

• Verbose explanations
• "You should" language
• Duplicate information
• Deep reference nesting
• README.md files