About Claude Skills Directory

This directory aggregates, evaluates, and ranks Claude skills, agents, and subagents from various community sources.

How It Works

  1. Aggregation: We collect skills from curated awesome lists like VoltAgent/awesome-claude-code-subagents

  2. Evaluation: Each skill is evaluated using an LLM-as-judge approach via OpenRouter, scoring:
    • Prompt Quality: Clarity, specificity, and adherence to best practices
    • Usefulness: Real-world applicability and completeness
  3. Metrics: We also display objective GitHub metrics (stars, forks, last commit) separately from quality ratings

Contributing

Found a skill that should be listed? Have feedback on evaluations? Visit our GitHub repository.

Data Sources

Currently indexing from:

More sources coming soon!