AI Tools¶

Effective and safe use of AI tools to improve development and documentation workflows.

AI tools are transforming software development. This section provides comprehensive guides for using AI assistants effectively and safely, with practical workflows and integration patterns.

🎯 Goals¶

Effective prompting patterns for development tasks
Safe practices to avoid security risks and maintain code quality
Workflow integration with existing development tools
Tool-specific guidance for each major AI platform

🔍 Quick Verification¶

Test your AI tool setup:

# Check Claude Code (if installed)
claude --version

# Check GitHub CLI for Copilot
gh --version
gh copilot --version

# Check if you have API access
# (Specific commands depend on tools you use)

# Verify environment variables for API keys
echo "API keys configured:" 
env | grep -E "(OPENAI|ANTHROPIC|GEMINI)" | cut -d= -f1

🤖 AI Tools Overview¶

Each tool has its strengths and ideal use cases. Choose the right tool for your specific needs:

📝 Claude Code ¶

Command-line AI assistant for development workflows
- Best for: File editing, code analysis, terminal tasks - Strengths: Deep codebase understanding, multi-file operations - Setup: Desktop app with command-line integration

💬 Claude Desktop ¶

Conversational AI for planning and problem-solving
- Best for: Architecture discussions, debugging complex issues - Strengths: Long-form conversations, visual content analysis - Setup: Desktop application with file uploads

🔧 ChatGPT ¶

Web-based AI assistant with extensive plugin ecosystem
- Best for: Quick questions, code explanations, learning - Strengths: Broad knowledge, fast responses, plugins - Setup: Web interface or API integration

🏠 Ollama ¶

Local AI models for privacy-sensitive work
- Best for: Offline development, sensitive codebases - Strengths: Complete privacy, customizable models - Setup: Local installation and model management

🌟 Gemini CLI ¶

Google's AI assistant with strong coding capabilities
- Best for: Code generation, multi-language projects - Strengths: Code understanding, integrated with Google services - Setup: CLI tool with API authentication

🚀 Common AI Workflows¶

Code Review & Analysis¶

Use Claude Code for automated file analysis
Use Claude Desktop for discussing architecture decisions
Use local models for sensitive code review

Learning & Documentation¶

Use ChatGPT for quick concept explanations
Use Claude Desktop for detailed documentation planning
Use AI tools to generate code examples and tutorials

Development Tasks¶

Code generation → Start with ChatGPT or Gemini for boilerplate
Debugging → Use Claude Code for file-level analysis
Refactoring → Use Claude Code for multi-file operations

🔐 Safety & Best Practices¶

Code Security¶

Never share production credentials, API keys, or sensitive data
Review all generated code before committing
Use local models (Ollama) for proprietary codebases
Set up proper .gitignore to avoid accidentally committing secrets

Quality Assurance¶

Test generated code thoroughly before deployment
Verify external dependencies and security implications
Maintain coding standards - AI should follow your conventions
Document AI-generated sections for future maintainers

Prompt Engineering¶

Be specific about requirements and constraints
Provide context about your codebase and goals
Iterate prompts to get better results
Save effective prompts for reuse

Python - Using AI tools with Python development
Git - AI-assisted Git workflows and commit messages
Toolkit - Integration with VS Code and other development tools

⚡ Quick References¶

Effective Prompting Patterns:

# Code Analysis
"Analyze this [language] file and identify potential issues with performance, security, and maintainability."

# Code Generation  
"Create a [language] function that [specific requirement]. Follow [coding standard] and include error handling."

# Debugging
"I'm getting [error message]. Here's the relevant code: [code]. What could be causing this issue?"

# Refactoring
"Refactor this function to be more readable and maintainable while preserving the exact same functionality."

Security Checklist: - [ ] No API keys or credentials in prompts - [ ] No production database URLs or sensitive endpoints - [ ] Review all generated code before committing
- [ ] Use local models for proprietary code - [ ] Check generated dependencies for security issues

Tool Selection Guide: - Quick questions → ChatGPT - File editing → Claude Code
- Architecture planning → Claude Desktop - Sensitive code → Ollama (local) - Code generation → Gemini CLI or ChatGPT - Long conversations → Claude Desktop

📁 Directory Structure¶

Each tool has its own directory with comprehensive documentation:

ai-tools/
├── index.md                 # This overview page
├── claude-code/
│   ├── index.md            # Claude Code introduction
│   ├── guides/             # Practical workflows
│   └── mcp/                # MCP configurations  
├── claude-desktop/
│   ├── index.md            # Claude Desktop setup
│   ├── guides/             # Usage patterns
│   └── mcp/                # MCP integrations
├── chatgpt/
│   └── index.md            # ChatGPT workflows
├── ollama/
│   ├── index.md            # Local model setup
│   └── models.md           # Model recommendations
└── gemini-cli/
    └── index.md            # Gemini CLI usage

Start here: If you're new to AI-assisted development, begin with Claude Code for practical development workflows, then explore Claude Desktop for planning and architecture discussions.