00-build-fine-tuning-skill

npx skills add majiayu000/claude-skill-registry --skill 00-build-fine-tuning-skill --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

# Build Your Fine-Tuning Skill You are about to learn supervised fine-tuning. But here is the pattern that separates effective AI-native developers from those who struggle: **build your skill first, then learn the technology**. In traditional learning, you study a topic, maybe take notes, and hope you remember it later. In Skill-First learning, you create a reusable intelligence asset before you even understand the technology deeply. This asset grows with you as you learn, and by the end of the chapter, you own a production-ready skill you can sell or deploy. This lesson follows the same pattern you used in Part 6 and Part 7. Clone a fresh skills-lab, fetch official documentation, and build your `llmops-fine-tuner` skill from authoritative sources rather than memory. ## Why Skill-First for Fine-Tuning? Fine-tuning has dozens of hyperparameters, multiple training strategies, and evolving best practices. Trying to memorize all of this is futile. But encoding it into a skill that you can invoke, test, and improve makes this knowledge permanently accessible and actionable. | Traditional Approach | Skill-First Approach | |---------------------|---------------------| | Read tutorials, fo

Clone the skills-lab repository
git clone https://github.com/panaversity/skills-lab.git ~/skills-lab-ch64
Navigate to the directory
cd ~/skills-lab-ch64
Create the skill directory structure
mkdir -p .claude/skills/llmops-fine-tuner
Create the learning spec
cat > LEARNING-SPEC.md << 'EOF'
cat > .claude/skills/llmops-fine-tuner/SKILL.md << 'EOF'
Check the skill exists