Installation Guide

This guide provides detailed instructions for installing the Claude Code Development Kit in your project.

Prerequisites

  • Required: Claude Code must be installed and accessible via the claude command in your terminal.
  • Recommended: For full functionality, you'll need MCP (Model Context Protocol) servers like Context7 for real-time library documentation and the Gemini Assistant for architectural consultations.

Platform Support

  • macOS / Linux: Fully supported.
  • Windows: Use via WSL (Windows Subsystem for Linux) is recommended. Native Windows support has reported bugs and is not officially supported.

This is the fastest and easiest way to get started. Run this single command in your terminal:

curl -fsSL https://raw.githubusercontent.com/peterkrueck/Claude-Code-Development-Kit/main/install.sh | bash

This script will: 1. Download the latest version of the kit. 2. Launch an interactive setup guide (setup.sh). 3. Ask you where to install the files (e.g., your project directory). 4. Prompt you to select optional components like MCP server integrations and audio notifications. 5. Copy the necessary files and generate a .claude/settings.local.json configuration file for you. 6. Provide links to install the optional MCP servers you selected.

https://github.com/user-attachments/assets/0b4a1e69-bddb-4b58-8de9-35f97919bf44

The installer is safe to run; it will detect any existing files and ask for permission before overwriting them.

Option 2: Clone and Install Manually

If you prefer to inspect the code before running it, you can clone the repository and run the setup script manually.

git clone https://github.com/peterkrueck/Claude-Code-Development-Kit.git
cd Claude-Code-Development-Kit
./setup.sh

This will launch the same interactive setup process as the quick install method.

What Gets Installed

The setup script creates the following structure in your target project directory:

your-project/
├── .claude/
│   ├── commands/          # AI orchestration templates (.md files)
│   ├── hooks/             # Automation scripts
│   │   ├── config/        # Security patterns configuration
│   │   ├── sounds/        # Notification sounds (if enabled)
│   │   └── *.sh           # Hook scripts (based on selections)
│   └── settings.local.json # Generated Claude Code configuration
├── docs/                  # Documentation templates and examples
│   ├── ai-context/        # Core documentation files
│   ├── open-issues/       # Issue tracking examples
│   ├── specs/             # Specification templates
│   ├── CONTEXT-tier2-component.md  # Component doc template
│   └── CONTEXT-tier3-feature.md    # Feature doc template
├── logs/                  # Hook execution logs (created at runtime)
├── CLAUDE.md              # Your project's master AI context
└── MCP-ASSISTANT-RULES.md # MCP coding standards (if Gemini selected)

Note: The exact files installed depend on your choices during the interactive setup.

Post-Installation Setup

After the script finishes, follow these steps to complete the setup:

  1. Customize Your AI Context:

    • Open CLAUDE.md and fill in your project's vision, architecture, and coding standards.
    • Open docs/ai-context/project-structure.md and document your project's tech stack and file structure.

  2. Install MCP Servers (if selected):

    • The installer will provide direct links to the repositories for any MCP servers you chose.
    • Follow their respective installation instructions.
    • Once installed and running, add their configuration to .claude/settings.local.json.

  3. Test Your Installation:

    • Navigate to your project directory and run claude.
    • Execute the following command to test context gathering: bash /full-context "analyze my project structure" Claude should read your new documentation and provide a summary.