everything-claude-code-zh  by xu-xiang

AI agent framework optimization system

Created 2 months ago
372 stars

Top 76.1% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

This repository offers a comprehensive, production-grade configuration suite for Claude Code, named "Everything Claude Code" (ECC). It provides agents, skills, hooks, commands, and rules, derived from a winning Anthropic hackathon project and extensive real-world usage. ECC aims to empower engineers, particularly Chinese speakers, with efficient tools for understanding and leveraging advanced AI agent capabilities.

How It Works

ECC functions as a modular system enhancing AI agent frameworks. Its core approach involves sophisticated token optimization, persistent memory via hooks, continuous learning to extract reusable skills from sessions, and robust validation loops. It supports parallelization and sub-agent orchestration, offering a complete ecosystem for building and deploying AI-driven applications. The configurations have evolved over months of high-intensity daily use.

Quick Start & Requirements

The recommended installation is via the Claude Code plugin: /plugin marketplace add affaan-m/everything-claude-code followed by /plugin install everything-claude-code@everything-claude-code. A manual installation script (./install.sh) is also available. Prerequisites include Claude Code CLI v2.1.0+ and Node.js for scripts.

Highlighted Details

  • AgentShield: An integrated security auditor with 1282 tests and 102 rules, scanning configurations for vulnerabilities.
  • Plankton: A write-time code quality enforcer using hooks and sub-processes for automated formatting and linting.
  • Continuous Learning v2: An instinct-based system that learns user patterns, scores confidence, and evolves them into skills.
  • Cross-Tool Compatibility: Maximizes functionality across Claude Code, Cursor IDE, Codex CLI, and OpenCode, featuring a DRY adapter pattern for Cursor.
  • Token Optimization: Provides recommended settings (sonnet model, reduced MAX_THINKING_TOKENS) and workflows to significantly lower operational costs.

Maintenance & Community

The project exhibits active development with frequent updates, including numerous community contributions (30+ PRs merged in v1.6.0). It is supported via GitHub Sponsors.

Licensing & Compatibility

Licensed under MIT, allowing free use, modification, and commercial integration. It is designed for broad compatibility with major AI agent frameworks like Claude Code, Cursor, Codex, and OpenCode.

Limitations & Caveats

Rules cannot be automatically distributed via the Claude Code plugin system, requiring manual installation. Over-enabling MCP servers can deplete the context window, necessitating careful management. The Codex CLI lacks hook support, relying on alternative security enforcement mechanisms.

Health Check
Last Commit

1 month ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
2
Star History
266 stars in the last 30 days

Explore Similar Projects

Feedback? Help us improve.