Your First Prompt¶

A step-by-step guide to creating and running your very first AIA prompt, from the simplest examples to more sophisticated patterns.

Before You Start¶

Make sure you have: - Installed AIA - Set up your API keys - Created your prompts directory (mkdir -p ~/.prompts)

The Simplest Prompt¶

Let's start with the most basic prompt possible:

Step 1: Create Your First Prompt File¶

echo "What is Ruby programming language?" > ~/.prompts/what_is_ruby.txt

Step 2: Run Your Prompt¶

aia what_is_ruby

That's it! AIA will: 1. Find your prompt file in ~/.prompts/ 2. Send it to the default AI model 3. Display the response in your terminal

Adding Context to Prompts¶

Now let's create a prompt that works with files:

Step 1: Create a Context-Aware Prompt¶

cat > ~/.prompts/explain_code.txt << 'EOF'
Please explain what this code does:

<%= context_file %>

Provide a clear explanation that covers:
- What the code accomplishes
- How it works
- Any potential improvements
EOF

Step 2: Use It with a File¶

# Create a sample code file
echo 'puts "Hello, World!"' > hello.rb

# Run the prompt with context
aia explain_code hello.rb

AIA automatically includes the file content where you specified <%= context_file %>.

Using Directives¶

Directives are special commands in prompts that start with //. Let's try some:

Configuration Directives¶

cat > ~/.prompts/detailed_analysis.txt << 'EOF'
//config model gpt-4
//config temperature 0.3

Provide a detailed technical analysis of this code:

<%= context_file %>

Focus on:
- Architecture and design patterns
- Performance considerations
- Security implications
- Best practices compliance
EOF

File Inclusion Directives¶

cat > ~/.prompts/project_overview.txt << 'EOF'
//include README.md
//include package.json

Based on the project files above, provide an overview of:
- Project purpose and goals
- Technology stack
- Getting started instructions
- Key dependencies
EOF

Shell Command Directives¶

cat > ~/.prompts/system_status.txt << 'EOF'
Current system status:

//shell date
//shell whoami
//shell uptime
//shell df -h

Please analyze the system status and provide recommendations.
EOF

Interactive Prompts with Parameters¶

Create prompts that accept parameters:

Step 1: Create a Parameterized Prompt¶

cat > ~/.prompts/code_review.txt << 'EOF'
//config model gpt-4
//config temperature 0.2

# Code Review: <%= file_name %>

Review this <%= language %> code for:
- Bugs and potential issues
- Code quality and style
- Performance optimizations
- Security considerations

Focus level: <%= focus_level || "standard" %>

Code to review:
//include <%= file_name %>

Please provide specific, actionable feedback.
EOF

Step 2: Use with Parameters¶

# ERB-style parameters (in prompt content)
aia code_review --file_name "app.py" --language "Python" --focus_level "security"

Your First Chat Session¶

Try AIA's interactive chat mode:

Step 1: Start a Chat¶

aia --chat

Step 2: Have a Conversation¶

You: Help me understand how to use AIA effectively
AI: I'd be happy to help! AIA is a powerful CLI tool for AI interactions...

You: Can you help me write a Python function?
AI: Of course! What kind of function would you like to create?

You: A function that calculates factorial
AI: Here's a Python factorial function...

Step 3: Save Your Conversation¶

You: /save factorial_help.md

Working with Different Models¶

Try different AI models for different tasks:

Quick Tasks¶

aia --model gpt-3.5-turbo what_is_ruby

Complex Analysis¶

aia --model gpt-4 detailed_analysis complex_code.py

Long Documents¶

aia --model claude-3-sonnet project_overview

Creating Your First Workflow¶

Chain prompts together for multi-step processes:

Step 1: Create Individual Steps¶

# Step 1: Extract requirements
cat > ~/.prompts/extract_requirements.txt << 'EOF'
//next analyze_requirements

Extract and list all requirements from this project:

//include README.md
//include requirements.txt

Provide a structured list of:
- Functional requirements
- Non-functional requirements
- Dependencies
- Constraints
EOF

# Step 2: Analyze requirements
cat > ~/.prompts/analyze_requirements.txt << 'EOF'
//next generate_recommendations

Based on the requirements extraction, analyze:
- Completeness of requirements
- Potential conflicts or gaps
- Implementation complexity
- Risk factors
EOF

# Step 3: Generate recommendations
cat > ~/.prompts/generate_recommendations.txt << 'EOF'
Based on the requirements and analysis, provide:
- Implementation recommendations
- Architecture suggestions
- Risk mitigation strategies
- Next steps
EOF

Step 2: Run the Workflow¶

aia extract_requirements

AIA will automatically run all three prompts in sequence!

Common Beginner Mistakes to Avoid¶

❌ Don't: Create overly complex first prompts¶

# Too complex for beginners
cat > ~/.prompts/bad_first_prompt.txt << 'EOF'
//config model gpt-4
//config temperature 0.7
//ruby complex_calculations
//shell complex_command | grep something
//include multiple_files.txt

Perform complex multi-step analysis with advanced features...
EOF

✅ Do: Start simple and build up¶

# Good first prompt
cat > ~/.prompts/good_first_prompt.txt << 'EOF'
Summarize this file in simple terms:

<%= context_file %>
EOF

❌ Don't: Ignore error messages¶

# If this fails, read the error message!
aia nonexistent_prompt

✅ Do: Use debug mode when learning¶

# See what AIA is doing
aia --debug --verbose your_prompt

Practice Exercises¶

Try these exercises to reinforce what you've learned:

Exercise 1: File Analyzer¶

Create a prompt that analyzes any file type and provides insights.

cat > ~/.prompts/analyze_file.txt << 'EOF'
Analyze this file:

//include <%= file %>

Provide:
- File type and format
- Key content summary
- Purpose and use case
- Any notable features
EOF

Exercise 2: Code Formatter¶

Create a prompt that suggests code improvements.

cat > ~/.prompts/improve_code.txt << 'EOF'
//config temperature 0.3

Suggest improvements for this code:

//include <%= code_file %>

Focus on:
- Readability
- Performance
- Best practices
- Error handling
EOF

Exercise 3: Project Documentation¶

Create a prompt that generates README content.

cat > ~/.prompts/generate_readme.txt << 'EOF'
Generate README.md content for this project:

Project structure:
//shell find . -type f -name "*.py" -o -name "*.rb" -o -name "*.js" | head -10

Configuration files:
//shell ls *.json *.yml *.yaml 2>/dev/null || echo "No config files found"

Create a comprehensive README with:
- Project description
- Installation instructions
- Usage examples
- Contributing guidelines
EOF

Tips for Success¶

Start Small¶

Begin with simple, single-purpose prompts
Test each prompt thoroughly before making it complex
Add one new feature at a time

Use Descriptive Names¶

# Good prompt names
~/.prompts/analyze_python_code.txt
~/.prompts/summarize_research_paper.txt
~/.prompts/generate_api_docs.txt

# Poor prompt names
~/.prompts/prompt1.txt
~/.prompts/test.txt
~/.prompts/stuff.txt

Organize Your Prompts¶

# Create categories
mkdir -p ~/.prompts/{development,analysis,writing,learning}

# Move prompts to appropriate directories
mv ~/.prompts/code_review.txt ~/.prompts/development/
mv ~/.prompts/analyze_file.txt ~/.prompts/analysis/

Version Control Your Prompts¶

cd ~/.prompts
git init
git add .
git commit -m "My first AIA prompts"

Experiment and Iterate¶

Try different models for the same prompt
Adjust temperature settings to see the difference
Refine prompts based on the results you get

Getting Help¶

Built-in Help¶

# General help
aia --help

# Model information
aia --available_models

# Debug information
aia --debug my_prompt

Community Resources¶

Check the FAQ for common questions
Browse Examples for inspiration
Read the Advanced Prompting guide when ready

Next Steps¶

After mastering your first prompt, explore:

Basic Usage - Common usage patterns
Chat Mode - Interactive conversations
Working with Models - Model selection strategies
Tools Integration - Extending capabilities
Advanced Prompting - Expert techniques

Troubleshooting Your First Prompt¶

Prompt Not Found¶

# Check if file exists
ls ~/.prompts/your_prompt.txt

# Check file permissions
chmod 644 ~/.prompts/your_prompt.txt

# Use full path if needed
aia --prompts_dir ~/.prompts your_prompt

API Errors¶

# Check API key
echo $OPENAI_API_KEY | cut -c1-10

# Test with simple prompt
aia --model gpt-3.5-turbo --debug simple_test

Unexpected Results¶

# Use debug mode
aia --debug --verbose your_prompt

# Try different model
aia --model claude-3-sonnet your_prompt

# Simplify the prompt
echo "Simple test question" | aia --chat

Congratulations!¶

You've created and run your first AIA prompt! You now understand: - ✅ How to create basic prompt files - ✅ How to run prompts with context - ✅ How to use simple directives - ✅ How to work with different models - ✅ How to start chat sessions - ✅ How to create basic workflows

Keep experimenting and building more sophisticated prompts as you become comfortable with these fundamentals!