Summary

This repository offers a comprehensive, battle-tested collection of Claude Code configurations—agents, skills, hooks, commands, rules, and MCPs—developed over 10+ months for production applications. It targets developers seeking to rapidly enhance their Claude Code environment with robust, reusable components, saving significant setup and refinement time.

How It Works

The project structures Claude Code functionality into modular components: agents for delegated sub-tasks, skills for workflows/domain knowledge, commands for direct invocation, hooks for event-driven automations, and rules for consistent guidelines. This approach facilitates organized, scalable AI-assisted development, leveraging specialized configurations for complex tasks.

Quick Start & Requirements

Clone the repo and copy desired configurations (agents, rules, commands, skills) into your ~/.claude/ directory. Integrate hooks by adding content from hooks/hooks.json to ~/.claude/settings.json, and configure MCP servers from mcp-configs/mcp-servers.json into ~/.claude.json, replacing placeholder API keys. Requires an active Claude Code setup. A detailed guide is linked in the README.

Highlighted Details

• Specialized Agents: Includes agents for planning, architecture, code review, security analysis, build error resolution, and E2E testing.
• Workflow Skills: Provides domain-specific skills for coding standards, backend/frontend patterns, TDD, and security checklists.
• Enforceable Rules: Comprehensive guidelines cover security, coding style, testing (TDD, 80% coverage), Git workflow, agent delegation, and performance.
• Event Hooks: Offers trigger-based automations for tool usage events.
• Production-Ready: Battle-tested configurations from a hackathon-winning project.

Maintenance & Community

Community contributions for new agents, skills, hooks, and configurations are encouraged via pull requests, with guidelines in CONTRIBUTING.md. Users can follow the author, @affaanmustafa, on X for updates.

Licensing & Compatibility

Released under the MIT License, permitting free use, modification, and distribution, including for commercial purposes. Users are encouraged to contribute back improvements.

Limitations & Caveats

A critical caveat concerns context window management: enabling too many MCPs/tools can severely degrade performance (e.g., 200k context window potentially reduced to 70k). Users must manage enabled tools per project (recommendation: <10 enabled, <80 total active) and utilize disabledMcpServers. Configurations require user adaptation for specific stacks.