Dynamic Development Environment Setup 🚀

╔═══════════════════════════════════════════════════════════╗
║                                                           ║
║     ███╗   ██╗███████╗██╗    ██╗    ██████╗ ███████╗██╗   ██╗
║     ████╗  ██║██╔════╝██║    ██║    ██╔══██╗██╔════╝██║   ██║
║     ██╔██╗ ██║█████╗  ██║ █╗ ██║    ██║  ██║█████╗  ██║   ██║
║     ██║╚██╗██║██╔══╝  ██║███╗██║    ██║  ██║██╔══╝  ╚██╗ ██╔╝
║     ██║ ╚████║███████╗╚███╔███╔╝    ██████╔╝███████╗ ╚████╔╝ 
║     ╚═╝  ╚═══╝╚══════╝ ╚══╝╚══╝     ╚═════╝ ╚══════╝  ╚═══╝  
║                                                           ║
║   ███████╗███████╗████████╗██╗   ██╗██████╗              ║
║   ██╔════╝██╔════╝╚══██╔══╝██║   ██║██╔══██╗             ║
║   ███████╗█████╗     ██║   ██║   ██║██████╔╝             ║
║   ╚════██║██╔══╝     ██║   ██║   ██║██╔═══╝              ║
║   ███████║███████╗   ██║   ╚██████╔╝██║                  ║
║   ╚══════╝╚══════╝   ╚═╝    ╚═════╝ ╚═╝                  ║
║                                                           ║
║       Dynamic Environment Auto-Configuration              ║
╚═══════════════════════════════════════════════════════════╝

Environment-aware setup system that automatically detects your context (VPS, Codespaces, local dev) and installs only the tools you need.

🎯 Smart Environment Detection

The setup script automatically detects your environment and configures it appropriately:

Environment	Auto-Detection	Tools Installed	Tools Skipped
VPS/Production	Hostname pattern, env vars	Tailscale, GitHub CLI, Doppler, AI assistant	None
GitHub Codespaces	CODESPACES env var	GitHub CLI, AI assistant, Doppler	Tailscale
Local Development	macOS + VS Code	All tools	None
CI/CD Pipeline	CI env vars	GitHub CLI	Others
Container/Docker	/.dockerenv file	GitHub CLI	Others

🛠️ Tools Included

• AI assistant (Codex CLI or Claude Code) - choose the agent that matches your subscription
• GitHub CLI (gh) - GitHub from the command line
• Tailscale - Zero-config VPN for secure networking
• Doppler - SecretOps platform for environment variables
• Vision MCP Server - Z.AI Vision capabilities for image and video analysis (when using Z.ai)
• Web Search MCP Server - Z.AI Search capabilities for real-time web search and information retrieval (when using Z.ai)
• Supabase MCP Server - Official Supabase integration for database management and queries (requires Supabase project)
• GitHub MCP Server - Official GitHub integration for repository intelligence and workflow automation (requires GitHub PAT)

🚀 Quick Start

For Z.ai Integration (feature branch)

# Clone the Z.ai integration branch
git clone -b feature/zai-integration https://github.com/ebowwa/seed.git
cd node-starter

# Quick setup (handles Doppler auth automatically)
./quick-setup.sh

For Standard Setup

# Clone and run - it auto-detects your environment!
git clone https://github.com/ebowwa/seed.git
cd node-starter
./setup.sh

📋 Prerequisites

• macOS: Command Line Tools (auto-prompts if missing)
• Linux: curl and sudo access
• All platforms: Internet connection

⚙️ Configuration

For Z.ai Integration

# The quick-setup.sh script handles everything automatically:
# 1. Installs Doppler CLI if needed
# 2. Prompts for Doppler authentication
# 3. Configures the project
# 4. Sets up Claude Code with Z.ai backend

./quick-setup.sh

Using Claude Code with Z.ai

After setup, you have two options to use Claude Code:

Option 1: Load environment per session

# Load Z.ai configuration from Doppler
eval $(doppler secrets download --config dev --format env --no-file)

# Start Claude Code with Z.ai backend
claude

Option 2: Use Doppler run wrapper

# Start Claude with Doppler environment
doppler run --project seed --config dev -- claude

Claude Settings Management

The project includes Claude settings templates for easy configuration:

# Install Z.ai-optimized Claude settings
./install-claude-settings.sh

# View settings files
ls .claude/
# .claude/settings.template.json  # Settings template
# .claude/settings.local.json     # Local project settings

The setup script automatically installs Claude settings based on your AI assistant preference:

• --use-zai installs Z.ai configuration (GLM-4.5 models)
• --use-claude installs standard Anthropic configuration
• The install-claude-settings.sh script is now dynamic and generates appropriate settings

Basic Setup (Standard)

# Optional: Add your API keys for automated config
cp .env.example .env
nano .env  # Add your tokens

# Run setup - auto-detects environment
./setup.sh

Force Specific Environment

# VPS/Production nodes
./setup.sh --env vps

# GitHub Codespaces
./setup.sh --env codespaces

# Local development
./setup.sh --env local_dev

# List all environments
./setup.sh --list-envs

Choose Your AI Assistant

Pick the agent that matches your subscription:

• claude (default): Claude Code with Anthropic’s API
• zai: Claude Code routed through the Z.ai Model API (GLM-4.5)
• codex: OpenAI Codex CLI

# Use Codex CLI instead of Claude Code
./setup.sh --assistant codex

# Route Claude Code through Z.ai’s endpoint
./setup.sh --assistant zai

# Shortcut flags
./setup.sh --use-codex
./setup.sh --use-claude
./setup.sh --use-zai

# Persist the preference via environment variable
echo "AI_ASSISTANT=zai" >> .env

The script defaults to Claude Code when no preference is set. At any time you can skip installing the selected assistant with --skip-assistant (aliases: --skip-claude, --skip-zai).

Skip Specific Tools

./setup.sh --skip-assistant   # Aliases: --skip-claude, --skip-zai
./setup.sh --skip-github
./setup.sh --skip-tailscale
./setup.sh --skip-doppler
./setup.sh --skip-vision      # Skip Vision MCP Server
./setup.sh --skip-search      # Skip Web Search MCP Server

# Combine options
./setup.sh --env vps --skip-doppler

🔑 Environment Variables

Create a .env file with your tokens for automated configuration:

# GitHub Personal Access Token
GITHUB_TOKEN=ghp_xxxxxxxxxxxx

# AI assistant preference (codex, claude, or zai)
AI_ASSISTANT=zai

# Z.ai API Key (required when AI_ASSISTANT=zai and for Vision/Web Search MCP Servers)
ZAI_API_KEY=zai-xxxxxxxxxxxx

# Anthropic API Key (optional when using Anthropic directly)
ANTHROPIC_API_KEY=sk-ant-xxxxxxxxxxxx

# Tailscale Auth Key (for headless setup)
TAILSCALE_AUTH_KEY=tskey-auth-xxxxxxxxxxxx

# Doppler Service Token
DOPPLER_TOKEN=dp.st.xxxxxxxxxxxx

🎨 Customization

Add Custom Environment

Edit situations.yaml:

environments:
  my_custom_env:
    description: "My special environment"
    detect:
      - env: "MY_ENV=true"
      - hostname_pattern: "custom-*"
    tools:
      - github-cli
      - doppler
    skip:
      - claude-code
      - tailscale

Add New Tools

Edit situations.yaml:

tools:
  my_tool:
    name: "My Tool"
    check_command: "mytool"
    install_methods:
      macos: "brew install mytool"
      debian: "apt-get install mytool"
    config_env: "MY_TOOL_TOKEN"

📚 Use Cases

VPS/Production Nodes

MACHINE_TYPE=vps ./setup.sh

• ✅ Tailscale for secure networking
• ✅ GitHub CLI for deployments
• ✅ Doppler for secrets management
• ✅ Preferred AI assistant (Codex, Claude, or Z.ai) for AI help

GitHub Codespaces

Automatically detected and configured:

• ✅ Preferred AI assistant (Codex, Claude, or Z.ai) for AI assistance
• ✅ GitHub CLI (essential)
• ✅ Doppler for dev secrets
• ❌ Tailscale (not needed)

Local Development

Full setup for your workstation:

• ✅ All tools installed
• ✅ Complete configuration

🔧 Post-Installation

Claude Code (if selected)

# Anthropic users
claude auth login

# Z.ai users
claude /status  # Confirms the GLM-4.5 endpoint is active

When AI_ASSISTANT=zai, the setup script writes your credentials to ~/.claude/settings.json (creating the file if needed) so future shells automatically target https://api.z.ai/api/anthropic.

Vision MCP Server (included with Z.ai)

When using Z.ai backend, the setup automatically configures Vision MCP Server for image and video analysis:

# Vision capabilities are available through Claude Code
# Place an image in your directory and ask Claude to analyze it:
claude
> What does this image show? (attach or reference image file)

# Vision MCP Server tools included:
# - image_analysis: Analyze images and provide detailed descriptions
# - video_analysis: Analyze videos and provide detailed descriptions

Web Search MCP Server (included with Z.ai)

When using Z.ai backend, the setup automatically configures Web Search MCP Server for real-time search capabilities:

# Search capabilities are available through Claude Code
# Ask Claude to search for current information:
claude
> Search for the latest AI technology developments

# Web Search MCP Server tools included:
# - webSearchPrime: Search web information with real-time results
#   * Page titles, URLs, and summaries
#   * Real-time news and information
#   * Stock prices, weather, and more

Supabase MCP Server (Manual Setup)

For users with Supabase projects, the official Supabase MCP server provides database management capabilities:

# Configure Supabase MCP server (manual setup required)
# See docs/Supabase_MCP_Integration.md for detailed instructions

# Requirements:
# - Supabase Personal Access Token (PAT)
# - Supabase project reference
# - Claude Code with MCP support

# Example usage after setup:
claude
> How many tables do we have in the database?
> Show me the structure of the apps table
> Insert a new record into the app_info table

Supabase MCP capabilities:

• Database schema exploration
• Natural language SQL queries
• Table and data management
• Direct database operations through Claude Code

GitHub MCP Server (Manual Setup)

For advanced GitHub operations, the official GitHub MCP server provides repository intelligence and workflow automation:

# Configure GitHub MCP server (manual setup required)
# See docs/GitHub_MCP_Integration.md for detailed instructions

# Requirements:
# - GitHub Personal Access Token (PAT)
# - Claude Code with MCP support

# Example usage after setup:
claude
> Find all TypeScript repositories in my organization
> Search for TODO comments across all my repositories
> Create issues for bugs found in the codebase
> Analyze commit patterns across my projects

GitHub MCP capabilities:

• Cross-repository code search
• Semantic code understanding
• Workflow automation (issues, PRs, projects)
• Organization-level analytics
• Natural language GitHub operations

Note: The seed project includes GitHub CLI for direct command operations. Use GitHub MCP for conversational tasks and CLI for precise control.

Codex CLI (if selected)

codex login

GitHub CLI

gh auth login

Tailscale

sudo tailscale up  # Linux/VPS
tailscale up       # macOS

Doppler

doppler login
# Or with service token:
doppler configure set token $DOPPLER_TOKEN --scope /

🐛 Troubleshooting

• Environment not detected? Use --env flag to force
• Tool already installed? Script checks and skips
• Permission errors? Ensure sudo access on Linux
• API keys not working? Check .env file formatting

📝 Files

• setup.sh - Main installation script
• quick-setup.sh - Streamlined Z.ai setup
• situations.yaml - Environment and tool configurations
• .env.example - Template for API keys
• .env - Your API keys (git-ignored)
• .claude/settings.template.json - Claude settings template for Z.ai
• .claude/settings.local.json - Local Claude settings
• install-claude-settings.sh - Claude settings installation script
• docs/Supabase_MCP_Integration.md - Complete Supabase MCP server setup guide
• docs/GitHub_MCP_Integration.md - Comprehensive GitHub MCP server documentation

🔒 Security

• Never commit .env files
• Use environment-specific tokens
• Rotate credentials regularly
• Review situations.yaml before running

📄 License

MIT - Use freely for your own setups!

🤝 Contributing

PRs welcome! Please test on target platforms.