cc-polymath: Context-Efficient Skills for Claude Code Development

Summary

cc-polymath is an innovative GitHub repository offering a vast collection of atomic skills and workflows specifically designed for Claude Code. It addresses the challenge of providing comprehensive development knowledge without overwhelming the AI's context window. By employing a progressive discovery system, cc-polymath ensures context efficiency and high-quality skill access across a wide range of technical domains.

Use at your own risk

OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.

Introduction

cc-polymath is an innovative GitHub repository that provides a comprehensive set of atomic skills and workflows for Claude Code. Its primary goal is to optimize context efficiency and skill quality, tackling the inherent tradeoff between providing extensive development knowledge and managing Claude's context window. This project achieves this by organizing over 410 skills through a multi-tier progressive discovery system.

Why Use and Benefits

The core problem cc-polymath solves is the "Context vs. Coverage" dilemma in AI-assisted development. Loading all necessary knowledge upfront can consume over 25,000 tokens, leaving little room for actual work, while loading nothing leaves Claude without essential patterns. cc-polymath's solution involves over 410 atomic, focused, and composable skills, each averaging 320 lines, ensuring quick loading and single responsibility.

The repository employs a three-tier architecture for progressive discovery:

Tier 1: Gateway Skills (23 auto-discovered Agent Skills): Lightweight entry points that activate automatically based on keywords in your prompt, such as discover-api or discover-database.
Tier 2: Category Indexes (30 detailed references): Full skill listings with descriptions and use cases, loaded only when browsing or planning.
Tier 3: Individual Skills (on-demand): Complete implementation guides loaded precisely when needed.

This architecture offers significant benefits:

Context efficiency: Reduces token consumption by 60-84% compared to monolithic indexes, freeing up context for code and conversation.
No boot overhead: Claude Code starts with zero skills loaded, with gateway skills activating intelligently.
Discoverability: Users don't need prior knowledge of available skills, as keywords in prompts trigger relevant gateways, enabling progressive loading from gateway to index to specific skill.

The repository covers a vast array of domains, including core development (languages, frameworks, infrastructure, DevOps, data, APIs), specialized domains (Machine Learning & AI, Rust & PyO3, Information Retrieval, Programming Language Theory, Formal Methods, Advanced Mathematics, Systems Programming), and workflow/meta skills.

Installation

Installation is straightforward, especially as a Claude Code Plugin.

# Recommended: Install as Claude Code Plugin
/plugin install https://github.com/rand/cc-polymath

# For local development or testing
/plugin install /Users/rand/src/cc-polymath

Once installed, all 410+ skills and the /skills command are immediately available. Auto-discovery works by simply starting your work, as Claude Code automatically activates relevant gateway skills based on your task keywords. Manual loading for browsing or planning is also supported using cat commands.

Examples

cc-polymath skills are designed to be composed for complex workflows. Here are a few examples:

Full-stack web app:

cat skills/api/rest-api-design.md
cat skills/frontend/nextjs-app-router.md
cat skills/database/postgres-schema-design.md
cat skills/caching/redis-caching-patterns.md
cat skills/cicd/github-actions-workflows.md

ML deployment with DSPy:

cat skills/ml/dspy-setup.md
cat skills/ml/dspy-modules.md
cat skills/ml/dspy-optimizers.md
cat skills/modal/modal-gpu-workloads.md
cat skills/modal/modal-web-endpoints.md
cat skills/observability/structured-logging.md

Multi-layer caching strategy:

cat skills/caching/caching-fundamentals.md
cat skills/caching/service-worker-caching.md
cat skills/caching/http-caching.md
cat skills/caching/cdn-edge-caching.md
cat skills/caching/redis-caching-patterns.md
cat skills/caching/cache-invalidation-strategies.md