Generated Skill

YAML
---
name: generating-ai-assistant-skills
description: Creates structured skill files for AI assistants by extracting methodologies from expertise descriptions. Use when you need to convert human expertise into reusable AI capabilities.
---

Python
def create_skill(expertise_description):
    """Convert expertise into structured skill format"""
    
    # Extract core methodology
    methodology = analyze_expertise(expertise_description)
    
    # Generate skill components
    skill = {
        'name': create_gerund_name(methodology.action),
        'description': craft_third_person_description(methodology),
        'quick_start': extract_immediate_example(methodology),
        'workflow': structure_process_steps(methodology),
        'examples': generate_input_output_pairs(methodology),
        'best_practices': distill_guidelines(methodology),
        'pitfalls': identify_common_mistakes(methodology)
    }
    
    return format_as_yaml_skill(skill)

1. Parse the expertise description

   • Identify the core professional activity
   • Extract key methodologies and processes
   • Note specific tools, frameworks, or approaches

2. Design the skill structure

   • Create gerund-form name (verb-ing format)
   • Write third-person description with trigger phrases
   • Structure content in the standard 5-section format

3. Generate actionable content

   • Start with immediate example (code/steps)
   • Build step-by-step workflow
   • Create concrete input/output examples
   • Distill best practices and pitfalls

4. Format and validate

   • Ensure YAML frontmatter compliance
   • Keep content concise (~500 lines max)
   • Verify all sections are practical and specific

Progress:

• Extract methodology from description
• Create compliant name and description
• Build Quick Start section
• Structure Workflow with steps
• Generate concrete Examples
• List Best Practices
• Identify Common Pitfalls
• Format as complete skill file

Example 1: Input: "I'm a data scientist who specializes in A/B testing. I design experiments, analyze results using statistical methods, and present findings to stakeholders." Output: Skill named conducting-ab-tests with sections covering experiment design, statistical analysis workflow, and stakeholder communication.

Example 2: Input: "Marketing copywriter focused on email campaigns. I write subject lines, body copy, and optimize for conversion rates." Output: Skill named writing-email-campaigns with quick-start templates, copywriting workflow, and conversion optimization techniques.

Example 3: Input: "Software architect who designs microservices. I break down monoliths, define service boundaries, and establish communication patterns." Output: Skill named designing-microservices with architecture patterns, service decomposition steps, and API design examples.

• Extract, don't invent - Base skills on actual described expertise
• Focus on methodology - Capture the "how" not just the "what"
• Make it actionable - Every section should enable immediate use
• Use concrete examples - Specific inputs/outputs beat abstract descriptions
• Keep it focused - One core expertise area per skill
• Include triggers - Help users know when to apply the skill
• Provide defaults - Give recommended approaches, not endless options

• Creating vague, generic skills that could apply to anything
• Using first or second person instead of third person in descriptions
• Omitting the "when to use" triggers in descriptions
• Making Quick Start too theoretical instead of immediately actionable
• Including too many options instead of recommended defaults
• Explaining basic concepts Claude already knows
• Creating workflows without clear step-by-step structure
• Using non-gerund or improperly formatted names
• Exceeding the practical length limit (~500 lines)