Skill Enhancer

Purpose

Transform existing Anthropic skills into enhanced, production-ready versions that are concise, well-structured, and optimally organized for Claude's progressive disclosure system.

When to Use This Skill

Use this skill when:

• User explicitly asks to improve, enhance, or optimize a skill
• Skill has quality issues (verbose, unclear structure, poor organization)
• User wants to upgrade legacy skills to current standards
• User mentions better skill effectiveness or token efficiency
• Skill files contain TODOs, second-person voice, or unused examples

Core Principles

1. Conciseness: Reduce SKILL.md to <5k words, move details to references/
2. Structure: Use clear patterns (workflow/task/reference/capabilities)
3. XML Integration: Mix natural text with XML tags for constraints, examples, requirements
4. Direct Instructions: Use imperative/infinitive voice (NOT second person)
5. Clean Output: Return enhanced skill ready for immediate use
6. Resource Optimization: Proper scripts/references/assets separation

Enhancement Process

Input Analysis

• Read complete skill (SKILL.md + all resources)
• Identify quality issues across dimensions:
  • Metadata quality (name, description, triggers)
  • Structural clarity (pattern, organization, flow)
  • Writing style (voice, conciseness, examples)
  • Resource organization (scripts, references, assets)
  • Size efficiency (token usage, progressive disclosure)

Core Enhancement Techniques

Metadata Enhancement

• Ensure kebab-case naming
• Write comprehensive description (30-60 words)
• Use third-person voice: "This skill should be used when..."
• Add specific triggers (file types, keywords, scenarios)

Voice Correction

• Eliminate second-person: "you should/can/will"
• Convert to imperative/infinitive: "Use X to...", "Process files by..."
• Remove filler words: "simply", "just", "basically", "obviously"
• Use active voice and direct instructions

Structural Optimization Apply appropriate pattern:

• Workflow-based: Sequential processes with decision trees
• Task-based: Multiple discrete operations
• Reference-based: Standards, guidelines, specifications
• Capabilities-based: Integrated systems with features

Content Refinement

• Add concrete examples (before/after transformations)
• Include XML tags for: constraints, requirements, examples, format, guidelines
• Keep main instructions as natural language
• Reference bundled resources with usage instructions

Resource Organization

• scripts/ - Executable code for repeated operations
• references/ - Detailed documentation (schemas, APIs, workflows)
• assets/ - Output templates, boilerplate, images
• Delete unused example files
• Eliminate duplication between SKILL.md and references

Size Reduction

• Move comprehensive docs to references/
• Add grep patterns for large reference files
• Implement progressive disclosure
• Target <5k words in SKILL.md

Output Format

Return the enhanced skill with:

• Updated SKILL.md with all improvements applied
• Reorganized resources (created/moved/deleted as needed)
• No meta-commentary about changes
• No TODO items remaining
• Immediately usable structure

Example Transformations

Example 1: Metadata Enhancement

Before:

description: Helps with PDFs

After:

description: Comprehensive PDF manipulation and analysis toolkit. This skill should be used when reading, creating, editing, merging, splitting, or extracting content from PDF files. Triggers include .pdf file paths, document manipulation requests, or form field operations.

Example 2: Voice Correction

Before:

You should use this function when you need to rotate a PDF.
When you're working with large files, you can optimize memory.

After:

Rotate PDFs using this function.
Optimize memory for large files.

Example 3: Structure with XML Integration

Before:

This skill helps analyze data. Use it for data analysis tasks.
It can handle various formats and provides insights.

After:

Analyze datasets and identify key insights, trends, and patterns.

<requirements>
- Highlight statistically significant findings
- Identify temporal trends and correlations
- Note anomalies or outliers
</requirements>

<format>
- Executive summary of findings
- Data visualizations as needed
- Detailed pattern analysis
- Actionable recommendations
</format>

Example 4: Resource Reorganization

Before:

skill-folder/
├── SKILL.md (8000 words with API docs, schemas, examples)
├── scripts/example.py (unused)
├── references/api_reference.md (empty)
└── assets/example_asset.txt (unused)

After:

skill-folder/
├── SKILL.md (2000 words, references resources)
├── scripts/
│   └── data_processor.py (actual utility)
└── references/
    ├── api_docs.md (detailed API reference)
    └── schemas.md (complete schemas)

Example 5: Progressive Disclosure

Before (all in SKILL.md):

# BigQuery Skill

## Complete API Reference
[2000 words of API documentation...]

## Database Schemas
[3000 words of schema details...]

## Query Examples
[2000 words of examples...]

After (SKILL.md + references):

# BigQuery Skill

## Purpose
Execute and analyze BigQuery queries with proper schema awareness.

## Quick Start
[300 words of essential patterns]

## Resources
<references>
Complete API documentation: references/api_docs.md
Database schemas: references/schemas.md

To find specific schema:
grep -i "table_name" references/schemas.md
</references>

Quality Standards