AI Agents - Setup & Configuration

Overview

Estimated time: 25โ€“35 minutes

This guide covers common setup patterns, authentication methods, and configuration best practices for AI coding assistants and agents.

Learning Objectives

Prerequisites

Authentication Patterns

API Key Based Authentication

Most AI coding tools require API keys for authentication:

OpenAI API Key

# Set environment variable
export OPENAI_API_KEY="sk-..."

# Or in .env file
OPENAI_API_KEY=sk-...

Used by: GitHub Copilot, Cursor, many CLI tools

Anthropic API Key

# Set environment variable
export ANTHROPIC_API_KEY="sk-ant-..."

# Configuration
ANTHROPIC_API_KEY=sk-ant-...

Used by: Claude integrations, Cursor

OAuth & Account Integration

# GitHub authentication for Copilot
gh auth login

# Verify authentication
gh auth status

# Replit account integration
replit auth login

Development Environment Setup

VS Code Configuration

Essential settings for AI-enhanced development:

{
  "editor.inlineSuggest.enabled": true,
  "editor.suggest.preview": true,
  "github.copilot.enable": {
    "*": true,
    "yaml": true,
    "plaintext": false,
    "markdown": true
  },
  "editor.quickSuggestions": {
    "comments": true,
    "strings": true,
    "other": true
  }
}

Terminal and Shell Setup

# Install helpful CLI tools
npm install -g @githubnext/github-copilot-cli

# Shell integration (add to .bashrc/.zshrc)
eval "$(github-copilot-cli alias -- "$0")"

# Warp terminal setup
curl -fsSL https://app.warp.dev/download | sh

Security Configuration

Security Best Practices:
  • Store API keys in environment variables, not in code
  • Use tool-specific credential management
  • Review data sharing policies for each tool
  • Configure local models when privacy is critical

Environment Variable Management

# Create .env file (add to .gitignore)
touch .env
echo ".env" >> .gitignore

# Use direnv for per-project environments
echo "export OPENAI_API_KEY=..." > .envrc
direnv allow

# Or use a credential manager
gh auth login --with-token < token.txt

Privacy-First Configuration

{
  "cursor.privacy.enableTelemetry": false,
  "cursor.ai.useLocalModels": true,
  "github.copilot.advanced": {
    "length": 500,
    "temperature": 0.1,
    "top_p": 1,
    "inlineSuggestCount": 3
  }
}

Tool-Specific Setup Guides

GitHub Copilot

# Install VS Code extension
code --install-extension GitHub.copilot

# CLI setup
npm install -g @githubnext/github-copilot-cli
echo 'eval "$(github-copilot-cli alias -- "$0")"' >> ~/.bashrc

Cursor Editor

# Download and install Cursor
curl -L https://cursor.sh/download -o cursor-installer

# Import VS Code settings
# File > Preferences > Import from VS Code

Replit Setup

# Install Replit CLI
npm install -g replit-cli

# Login and configure
replit auth login
replit whoami

Configuration Files

Project-Level Configuration

Create standardized configuration files for consistent AI assistance:

# .ai-config.yml
tools:
  github-copilot:
    enabled: true
    suggestions: inline
  cursor:
    composer: true
    chat: true
  
privacy:
  data_sharing: false
  local_models: preferred
  
preferences:
  code_style: consistent
  documentation: auto-generate

Workspace Settings

{
  "ai.enabled": true,
  "ai.suggestions.enabled": true,
  "ai.chat.enabled": true,
  "editor.formatOnSave": true,
  "editor.codeActionsOnSave": {
    "source.organizeImports": true,
    "source.fixAll": true
  }
}

Troubleshooting Common Issues

Authentication Problems

GitHub Copilot not working

# Check authentication
gh auth status

# Re-authenticate
gh auth login --web

# Check Copilot status
gh copilot status

API rate limits

  • Check your API usage in provider dashboards
  • Implement request throttling
  • Use caching for repeated requests
Performance Issues

Slow suggestions

  • Reduce suggestion count in settings
  • Clear editor cache and restart
  • Check network connectivity
  • Use local models for better performance

High resource usage

{
  "ai.maxTokens": 1000,
  "ai.requestTimeout": 5000,
  "ai.cacheDuration": 300
}

Best Practices

Security

  • Never commit API keys
  • Use project-specific credentials
  • Regular key rotation
  • Monitor usage patterns

Performance

  • Configure appropriate timeouts
  • Use caching strategically
  • Optimize suggestion frequency
  • Monitor resource usage

Workflow

  • Consistent configuration across team
  • Document setup procedures
  • Regular tool updates
  • Backup configurations

Privacy

  • Review data policies
  • Use local models when possible
  • Configure telemetry settings
  • Audit data sharing

Next Steps

With your development environment configured, you're ready to explore specific AI tools: