Stand on the shoulders of giants — Create AI skills built on proven methodologies from domain experts.

A skill that helps you discover and incorporate frameworks, principles, and best practices from recognized masters before generating any new skill. Works with Claude Code, Codex, and other AI agent platforms.

The hard part of creating a skill isn't the format — it's knowing the best way to do the thing.

Most professional domains have masters who spent decades figuring out what works:

• Jobs on product, hiring, and marketing
• Bezos on writing (6-pager) and decision-making
• Munger on mental models
• Chris Voss on negotiation

This skill surfaces their methodologies before you write a single line, so your skill embodies world-class expertise from day one.

1. You: "I want to create a skill for user interviews"

2. Skill-from-masters:
   ├── Checks local methodology database
   ├── Searches web for additional experts
   ├── Finds golden examples of great outputs
   ├── Identifies common mistakes to avoid
   └── Cross-validates across sources

3. Surfaces experts:
   - Rob Fitzpatrick (The Mom Test)
   - Steve Portigal (Interviewing Users)
   - Nielsen Norman Group best practices

4. You select which methodologies to incorporate

5. Extracts actionable principles from primary sources

6. Hands off to skill-creator to generate the final skill

The skill includes a curated database covering 15+ domains:

Plus an "Oral Tradition" section for experts who share primarily through talks and interviews (Steve Jobs, Elon Musk, Jensen Huang, Patrick Collison, etc.)

# Navigate to your skills directory
cd ~/.claude/skills  # Claude Code
# or your platform's skills path

# Clone this skill
git clone https://github.com/anthropics/skill-from-masters.git

1. Download the skill-from-masters folder
2. Place it in your skills directory

This skill works with skill-creator. Make sure skill-creator is available in your environment.

The skill automatically activates when you ask to create a new skill. It runs before skill-creator to ensure your skill is built on proven methodologies.

You: "Help me create a skill for writing product requirement documents"

Skill-from-masters:

Let me search for the best methodologies on PRD writing.

[Searches: "PRD best practices expert", "product requirements document methodology"]

I found several strong frameworks:

Product Management:

• Marty Cagan: Problem-first PRDs, separating problem space from solution space
• Teresa Torres: Outcome-focused requirements

Business Writing:

• Barbara Minto: Pyramid Principle - lead with conclusion, MECE structure
• Amazon: Working backwards, 6-pager narrative format

Golden Examples Found:

• Intercom's public PRD template
• Linear's product spec format

Common Mistakes:

• Jumping to solutions before defining the problem
• Writing for engineers instead of for alignment
• Missing success metrics

Which approaches resonate? Or should I search deeper into any of them?

You: "I like Amazon's approach combined with Cagan's problem-first thinking."

Skill-from-masters: [Fetches primary sources on Amazon's methodology, extracts principles, then generates skill via skill-creator]

This repository contains the following skills:

When you want to create a new skill based on expert methodologies:

• 3-layer search: local database → web experts → primary sources
• Finds golden examples and anti-patterns
• Cross-validates across multiple experts
• Hands off to skill-creator for final generation

Example:

You: "Help me create a skill for user interviews"
→ Finds: Rob Fitzpatrick (The Mom Test), Steve Portigal, Nielsen Norman Group
→ You select which methodologies to incorporate
→ Generates skill with those principles encoded

When you want to find an existing skill instead of creating one:

• Searches only 5 trusted sources (no random internet results)
• Tier-based priority: official → curated → aggregators
• Filters out low-quality results (stars < 10, outdated, no SKILL.md)
• Security checks for suspicious code patterns

Example:

You: "I need a skill for frontend design, automated testing, and code review"
→ Searches: anthropics/skills, ComposioHQ, travisvn, skills.sh, skillsmp.com
→ Returns: frontend-design (official), webapp-testing (official), code-review-excellence (26k stars)

When you want to learn from a GitHub project and turn that knowledge into a skill:

• Search GitHub for quality projects (stars > 100, actively maintained)
• Present options and wait for your confirmation
• Deep dive into selected project (README, source code, examples)
• Summarize what it learned, then create skill via skill-creator

Example:

You: "I want to convert images to ASCII art"
→ Searches GitHub, finds: ascii-image-converter (3.1k stars), RASCII (224 stars)
→ You select ascii-image-converter
→ Learns: brightness-to-character mapping, aspect ratio handling, color techniques
→ Creates skill encoding that knowledge (not just wrapping the tool)

Key difference: This skill encodes the knowledge from projects, so the skill works even without the original tool installed.

skill-from-masters/
├── skill-from-masters/
│   ├── SKILL.md                              # Core skill: create from expert methodologies
│   └── references/
│       ├── methodology-database.md           # Curated expert frameworks
│       └── skill-taxonomy.md                 # 11 skill type categories
├── skills/
│   ├── search-skill/
│   │   └── SKILL.md                          # Search existing skills from trusted sources
│   └── skill-from-github/
│       └── SKILL.md                          # Learn from GitHub projects
├── README.md
├── LICENSE
└── .gitignore

Before finalizing any skill, this skill verifies:

• Searched beyond the local database
• Found primary sources, not just summaries
• Found golden examples of the output
• Identified common mistakes to avoid
• Cross-validated across multiple experts
• Encoded specific, actionable steps (not vague principles)

Contributions welcome! Especially:

• Adding new domains and experts to the methodology database
• Improving framework descriptions with source links
• Sharing examples of skills created with this approach

Please:

1. Fork the repository
2. Create a feature branch
3. Submit a pull request

MIT License — feel free to use, modify, and distribute.

Philosophy: Quality isn't written. It's selected.