Skene Cookbook Metrics

Skene Cookbook Metrics

Last Updated: 2026-02-12 Version: 0.1.1 Canonical Source of Truth — All documentation must reference this file for accuracy

Official Numbers

Skill Chain Cookbook

• Ready-to-use recipes: 36 (Skill Chains — the primary product)
• Each recipe chains 2–7 skills from the library below; see SKILL_CHAINS.md.

Skill Counts (ingredients for recipes)

• Total Skills: 764
• Executable Skills: 382
• Reference Guides: 382
  • Cursor Rules: 241
  • Scientific Computing: 141

Domain Counts

• Executable Domains: 21
• Reference Domains: 2
• Total Unique Domains: 23

Executable Skills Breakdown

By Domain

Total Executable: 382 skills across 21 domains

By Job Function

Mapping domains to typical job functions:

Note: Some domains serve multiple functions. Total exceeds 382 due to overlapping categorization.

Reference Guides Breakdown

Cursor Rules (241 skills)

Technology-specific coding guidelines and best practices for Cursor IDE:

• Web frameworks (React, Vue, Angular, etc.)
• Backend frameworks (Flask, Django, Express, etc.)
• Cloud platforms (AWS, GCP, Azure)
• Databases (PostgreSQL, MongoDB, Redis, etc.)
• DevOps tools (Docker, Kubernetes, Terraform, etc.)
• Languages (Python, JavaScript, TypeScript, Go, Rust, etc.)
• And 200+ more specialized tools and frameworks

Location: skills-library/reference/cursor_rules/

Scientific Computing (141 skills)

Domain-specific knowledge for research and scientific workflows:

• Bioinformatics tools (BLAST, UniProt, PDB, etc.)
• Data analysis libraries (NumPy, Pandas, SciPy, etc.)
• Visualization tools (Matplotlib, Seaborn, Plotly, etc.)
• Machine learning frameworks (scikit-learn, TensorFlow, PyTorch, etc.)
• Research databases and APIs
• Academic workflow tools

Location: skills-library/reference/scientific/

Counting Methodology

What Counts as "One Skill"?

A skill is defined by the presence of a skill.json file in a skill directory.

Example skill directory structure:

skills-library/executable/marketing/seo-optimizer/
├── skill.json          ← This defines the skill
├── metadata.yaml       ← Additional metadata
├── instructions.md     ← Human-readable docs
└── references/         ← Optional reference materials
    ├── api_docs.md
    └── examples.md

Counting Command:

# Count executable skills
find skills-library/executable -name "skill.json" | wc -l

# Count reference guides
find skills-library/reference -name "skill.json" | wc -l

# Count all skills
find skills-library -name "skill.json" | wc -l

Executable Skills vs Reference Guides

Executable Skills (skills-library/executable/):

• AI agent skills with tools and execution logic
• Invokable by Claude/Cursor to perform actions
• Include tool definitions, parameters, error handling
• Production-ready with security gates
• Count: 382 skills

Reference Guides (skills-library/reference/):

• Reference materials and domain knowledge
• Cursor rules for IDE-specific guidance
• Scientific computing documentation
• Not directly executable, but provide reference knowledge for AI
• Count: 382 skills

Why Multiple Numbers?

You may see different numbers referenced in different places:

All of these are correct for their respective audiences. When in doubt:

• User-facing claims: Use "760+ skills" (total)
• Technical accuracy: Use "382 executable + 382 reference guides = 764 total"
• Marketing copy: Use "760+ skills across 23 domains"

Single Source of Truth

For All Documentation

When writing or updating documentation:

1. Reference this file for canonical numbers
2. Link to METRICS.md rather than hardcoding numbers
3. Use consistent terminology:
   • "760+ skills" or "764 total skills" for general use
   • "382 executable skills" for technical precision
   • "23 domains" for domain count (21 executable + 2 reference)

Verification

To verify these numbers match the actual repository:

# Navigate to repository
cd /path/to/skene-cookbook

# Count executable skills by domain
for dir in skills-library/executable/*/; do
  echo "$(basename "$dir"): $(find "$dir" -name 'skill.json' | wc -l | tr -d ' ')"
done

# Count reference guides
echo "cursor_rules: $(find skills-library/reference/cursor_rules -name 'skill.json' | wc -l | tr -d ' ')"
echo "scientific: $(find skills-library/reference/scientific -name 'skill.json' | wc -l | tr -d ' ')"

# Count totals
echo "Total executable: $(find skills-library/executable -name 'skill.json' | wc -l | tr -d ' ')"
echo "Total reference: $(find skills-library/reference -name 'skill.json' | wc -l | tr -d ' ')"
echo "Grand total: $(find skills-library -name 'skill.json' | wc -l | tr -d ' ')"

Automated Verification

A verification script ensures documentation stays aligned:

# Run verification
npm run verify:metrics

# This checks all documentation files against actual filesystem counts
# Fails with exit code 1 if any mismatches are found

See scripts/sync_metrics.js for implementation details.

Historical Context

Why This File Exists

Prior to v0.1.2, skill counts were inconsistent across documentation:

• README.md claimed "760+ skills"
• index.json showed different numbers
• Various docs had conflicting domain counts
• No clear definition of "what is a skill"

This file solves that problem by establishing:

1. Clear counting methodology
2. Single canonical source
3. Automated verification
4. Consistent terminology across all docs

Version History

Note: Counts for 0.1.0 and 0.1.1 are retroactive estimates based on git history.

Maintenance

When Skills Are Added/Removed

1. Update METRICS.md first with new counts
2. Run verification script to ensure accuracy:

   npm run verify:metrics
   

3. Update other docs if categories/domains change
4. Commit with clear message indicating metric changes

Quarterly Reviews

Every quarter, verify:

• Skill counts match filesystem
• Domain categorization is accurate
• Job function mappings are current
• All documentation references METRICS.md
• Verification script is working

Quick Reference

For Copy/Paste:

Total Skills: 764
Executable: 382 (across 21 domains)
Reference: 382 (cursor_rules: 241, scientific: 141)
Total Domains: 23

Marketing Tagline:

"760+ AI skills for Claude and Cursor across 23 domains"

Technical Description:

"382 executable AI agent skills + 382 reference guides = 764 total resources"

Domain Breakdown:

"21 executable domains (marketing, PLG, RevOps, etc.) + 2 reference domains (cursor_rules, scientific)"

Links

• README.md — Main repository documentation
• skills-library/README.md — Skills library overview
• SKILLS_CATALOG.md — Complete skill catalog
• index.json — Machine-readable skill index
• CHANGELOG.md — Version history

This file is the canonical source of truth for all skill metrics. When in doubt, reference this file.