Everything Claude Code: The Ultimate Performance System for AI Agent Harnesses

An Anthropic hackathon winner that grew into the most popular AI agent enhancement system—now with 174K+ GitHub stars.

What Is Everything Claude Code?

#

Everything Claude Code (ECC) is an agent harness performance optimization system—a complete enhancement layer for AI coding agents like Claude Code, Codex, Cursor, OpenCode, and Gemini.

It started as a personal setup by @affaan-m, who won the Anthropic x Forum Ventures hackathon building zenith.chat entirely with Claude Code. After 10+ months of daily use building real products, he open-sourced his entire workflow. The result? A comprehensive system that has since accumulated 174,000+ stars on GitHub and over 28,000 forks.

ECC is not just a collection of config files. It’s a complete, production-ready system:

• 300+ skills — reusable prompt bundles for everything from code review to TDD to research
• Instincts — automatic behaviors that guide Claude’s default responses
• Memory optimization — persistent knowledge graph for long-running projects
• Continuous learning — automated improvement loops that refine outputs over time
• Security scanning — built-in vulnerability detection for every commit
• Research-first development — automated multi-source research before code changes
• MCP configurations — ready-to-use Model Context Protocol server setups

And it’s completely free and MIT-licensed.

Why ECC Exists

#

If you’ve used Claude Code, you know it’s powerful out of the box. But as you use it daily across real projects, you realize:

• You keep writing the same instructions
• Context windows fill up with setup boilerplate
• You spend time re-teaching Claude your conventions
• There’s no built-in system for continuous improvement

ECC solves all of this by providing a structured enhancement layer that sits on top of your agent harness. Install it once, and suddenly your AI agent has:

• A complete library of specialized skills at its disposal
• Automatic memory of project conventions and decisions
• Built-in quality gates for security and code review
• Research capabilities that kick in before writing code

Key Features

#

Skills System

#

Skills are the primary workflow surface in ECC. They act like scoped workflow bundles: reusable prompts, structure, supporting files, and codemaps that let Claude quickly navigate your codebase without burning context on exploration.

Instead of describing what you want every time, you invoke a slash command:

/tdd          → Test-driven development workflow
/code-review  → Comprehensive code review
/e2e          → End-to-end testing setup
/refactor-clean → Dead code cleanup
/security-review → Security vulnerability scan
/deep-research → Multi-source research with synthesis

Each skill includes structured prompts, command handlers, and codemaps—everything Claude needs to execute that workflow correctly from the start.

Instincts

#

Instincts are automatic behaviors that guide how Claude responds by default. Instead of starting from scratch each session, ECC configures instincts for:

• Code quality standards (immutability, error handling, naming conventions)
• Project-specific conventions (file organization, testing patterns)
• Security-first development (input validation, secret detection)
• Documentation discipline (when and how to document)

Memory Optimization

#

ECC implements a persistent knowledge graph that tracks project entities, decisions, and relationships across sessions. This means your AI agent remembers:

• Architecture decisions and their rationale
• Which patterns work for your specific project
• Key entities and their relationships
• Past decisions that should inform future work

No more re-teaching Claude everything every session.

Continuous Learning

#

ECC includes feedback loops that automatically improve outputs over time:

• Run evaluations to measure quality
• Detect patterns in past successes and failures
• Refine prompts and instincts based on results
• Build up a custom knowledge base for your domain

Security Scanning

#

Before every commit, ECC can run a security scan that checks for:

• Hardcoded secrets and API keys
• SQL injection vulnerabilities
• XSS and CSRF patterns
• Insecure cryptographic usage
• Dependency vulnerabilities

Research-First Development

#

ECC promotes a research-before-coding workflow. When you ask it to implement a feature, it automatically:

1. Searches existing code and GitHub for patterns
2. Checks library documentation via Context7
3. Synthesizes findings into a plan
4. Then—and only then—starts writing code

Cross-Harness Compatibility

#

One of ECC’s standout features is that it works across multiple AI agent harnesses. You’re not locked into a single tool:

Install once, use everywhere.

What’s Inside

#

ECC contains a massive collection of production-tested components:

• 300+ skills covering development, testing, security, research, and operations
• 100+ agents for specialized tasks (code review, security, planning, research)
• MCP server configurations for Context7, Exa search, databases, GitHub, and more
• Hooks system for PostToolUse formatting, linting, and type checking
• Language-specific rules for TypeScript, Python, Go, Rust, Swift, PHP, and more
• VSCode integration and editor configurations
• Plugin system for seamless Claude Code integration
• Template projects for rapid bootstrapping

The project spans multiple languages—JavaScript (57%), Rust (33%), Python (5%), Shell (3%), and TypeScript (1%)—with 170+ contributors.

How to Install

#

Option 1: Install as Plugin (Recommended)

#

The easiest way is to install ECC as a Claude Code plugin:

# Add the marketplace
/plugin marketplace add https://github.com/affaan-m/everything-claude-code

# Install the plugin
/plugin install ecc@ecc

That’s it. You now have access to all commands, agents, skills, and hooks.

Option 2: Manual Installation

#

Clone the repo and copy what you need:

git clone https://github.com/affaan-m/everything-claude-code.git
cd everything-claude-code

# Copy rules for your tech stack (e.g., TypeScript + Web)
cp -r rules/typescript ~/.claude/rules/
cp -r rules/web ~/.claude/rules/
cp -r rules/common ~/.claude/rules/

# Copy skills you want
cp -r skills/* ~/.claude/skills/

Option 3: Via settings.json

#

Add directly to your Claude configuration:

{
  "extraKnownMarketplaces": {
    "ecc": {
      "source": {
        "source": "github",
        "repo": "affaan-m/everything-claude-code"
      }
    }
  },
  "enabledPlugins": {
    "ecc@ecc": true
  }
}

Getting Started

#

Once installed, start with these commands:

/hookify        → Set up automation hooks for your project
/init           → Initialize ECC for your project
/tdd            → Start a test-driven development workflow
/code-review    → Review your current changes
/deep-research  → Research before implementing

Or dive into the shorthand guide: /projects shows project-level commands, /plan for structured feature planning.

Why It Matters

#

AI coding agents are getting better at an astonishing rate, but raw model capability is only half the equation. The other half is how you direct it.

ECC provides the structure, memory, and workflow that transforms an AI agent from a clever autocomplete into a reliable, consistent development partner. It’s the difference between asking Claude to “review this code” and having it check 15 specific quality gates across security, style, testing, and architecture.

With 174K+ stars and a thriving community of 170+ contributors, ECC has become the de facto standard for enhancing AI coding agents. It’s not just a tool—it’s a movement toward agentic development done right.

Learn More:

• GitHub: github.com/affaan-m/everything-claude-code
• Website: ecc.tools
• License: MIT — free to use, modify, and share