Skip to content

Prompt Examples

This collection contains real-world prompt examples organized by category and complexity level.

Categories

Development

Prompts for software development tasks: - Code Review: Quality analysis and improvement suggestions - Documentation: Generate comprehensive code documentation - Debugging: Systematic problem diagnosis and resolution - Architecture: System design analysis and recommendations - Testing: Test strategy and implementation guidance

Writing

Content creation and editing prompts: - Technical Writing: API docs, tutorials, technical guides - Blog Posts: Engaging technical and general content - Creative Writing: Stories, poetry, creative projects - Editing: Content improvement and style refinement - Marketing: Copy, descriptions, promotional content

Analysis

Data analysis and research prompts: - Data Analysis: Statistical analysis and insights - Research: Literature review and synthesis - Reports: Structured analysis and recommendations - Comparison: Competitive analysis and evaluation - Trends: Pattern recognition and forecasting

Automation

System administration and automation prompts: - System Monitoring: Health checks and diagnostics - Deployment: Release and deployment workflows - Log Analysis: System log interpretation - Maintenance: Routine system maintenance tasks - Alerting: Notification and response templates

Learning

Educational and knowledge acquisition prompts: - Concept Explanation: Complex topic simplification - Tutorial Creation: Step-by-step learning guides - Quiz Generation: Assessment and evaluation tools - Research Assistance: Academic and professional research - Skill Development: Practice exercises and challenges

Complexity Levels

Basic

  • Simple, single-purpose prompts
  • Minimal configuration required
  • Clear, straightforward outputs
  • Great for learning AIA basics

Intermediate

  • Multi-step workflows
  • Dynamic configuration
  • Context-aware processing
  • Suitable for regular use

Advanced

  • Complex multi-stage pipelines
  • Extensive use of directives
  • Tool and MCP integration
  • Production-ready workflows

Using These Examples

1. Copy to Your Prompts Directory

# Copy individual prompts
cp docs/examples/prompts/development/code_review.txt ~/.prompts/

# Copy entire categories
cp -r docs/examples/prompts/development/ ~/.prompts/

# Copy all examples
cp -r docs/examples/prompts/* ~/.prompts/

2. Customize for Your Needs

Each prompt includes customization sections: - Parameters: Variables you can adjust - Configuration: Settings to modify - Extensions: How to add functionality - Variations: Alternative approaches

3. Run Examples

# Basic usage
aia code_review my_file.py

# With customization
aia --model gpt-4 --temperature 0.3 code_review my_file.py

# In workflows
aia --pipeline "code_review,optimize,test" my_project/

Code Review Prompt

File: development/code_review.txt 

//config model gpt-4
//config temperature 0.3

# Code Review Analysis

Review the following code for:
- **Bugs**: Logic errors, edge cases, potential crashes
- **Security**: Vulnerabilities, input validation, data exposure
- **Performance**: Efficiency, scalability, resource usage
- **Style**: Conventions, readability, maintainability
- **Best Practices**: Design patterns, industry standards

## Code to Review:
//include <%= file %>

## Review Format:
Provide your analysis in the following structure:

### Summary
Brief overall assessment and rating (1-10).

### Issues Found
List specific problems with severity levels:
- 🔴 **Critical**: Security vulnerabilities, crashes
- 🟠 **Major**: Performance issues, bugs
- 🟡 **Minor**: Style, minor improvements

### Recommendations
Concrete suggestions for improvement with code examples where applicable.

### Positive Aspects
Highlight what's done well in the code.

Blog Post Generator

File: writing/blog_post.txt 

//config model gpt-4
//config temperature 1.0
//config max_tokens 3000

# Technical Blog Post Generator

Create an engaging, well-structured blog post about: **<%= topic %>**

## Requirements:
- **Target Audience**: <%= audience || "Software developers" %>
- **Word Count**: <%= word_count || "1000-1500 words" %>
- **Tone**: <%= tone || "Professional but approachable" %>
- **Include Code Examples**: <%= code_examples || "Yes" %>

## Context:
<% if context_file %>
//include <%= context_file %>
<% end %>

## Structure:
1. **Hook**: Engaging opening that grabs attention
2. **Introduction**: Problem statement and article overview
3. **Main Content**: 3-4 major sections with headers
4. **Code Examples**: Practical, runnable code samples
5. **Best Practices**: Key takeaways and recommendations
6. **Conclusion**: Summary and call-to-action

## Style Guidelines:
- Use clear, concise language
- Include practical examples
- Add subheadings for readability
- Include relevant links and resources
- End with actionable next steps

Please ensure the post is SEO-friendly with good header structure and includes relevant keywords naturally.

Data Analysis Workflow

File: analysis/data_pipeline.txt 

//config model claude-3-sonnet
//config temperature 0.2

# Data Analysis Pipeline

Analyze the provided dataset and generate comprehensive insights.

## Dataset Information:
//shell head -5 <%= dataset_file %>
//shell wc -l <%= dataset_file %>
//shell file <%= dataset_file %>

## Analysis Steps:

### 1. Data Overview
- Examine data structure and types
- Identify columns and their meanings
- Note data quality issues

### 2. Descriptive Statistics
- Calculate summary statistics
- Identify distributions and outliers
- Examine correlations

### 3. Data Quality Assessment
- Missing values analysis
- Duplicate detection
- Inconsistency identification

### 4. Key Insights
- Significant patterns and trends
- Interesting correlations
- Anomalies or outliers

### 5. Recommendations
- Data cleaning suggestions
- Further analysis opportunities
- Actionable business insights

## Data Sample:
//include <%= dataset_file %>

Please provide a thorough analysis with specific findings and quantitative metrics where possible.

Prompt Design Patterns

Parameterization Pattern

Make prompts reusable with variables: 

//config model <%= model || "gpt-4" %>
//config temperature <%= temperature || "0.7" %>

Task: <%= task_description %>
Context: <%= context || "General" %>
Output Format: <%= format || "Markdown" %>

Conditional Inclusion Pattern

Include different content based on conditions: 

<% if File.exist?('production.yml') %>
//include production.yml
<% else %>
//include development.yml
<% end %>

<% if ENV['DETAILED_ANALYSIS'] == 'true' %>
Provide detailed technical analysis.
<% else %>
Provide summary analysis.
<% end %>

Multi-Stage Pipeline Pattern

Chain related prompts together: 

//next data_cleaning
//pipeline analysis,visualization,reporting

Initial data processing completed.
Ready for next stage: <%= next_stage %>

Tool Integration Pattern

Incorporate external tools: 

# Get a list of tools that are available
//tools

Using advanced analysis tools:

# Tell the LLM which tool to use and its arguments
use the examine_data tool to review this file '<%= data_file %>')

Validation and Testing

Testing Your Prompts

  1. Syntax Check: Verify directive syntax
  2. Parameter Testing: Test with different inputs
  3. Output Validation: Ensure consistent, quality outputs
  4. Performance Testing: Check response times and costs
  5. Edge Case Testing: Handle unusual inputs gracefully

Example Test Scripts

# Test basic functionality
aia --debug code_review test_file.py

# Test with different models
for model in gpt-3.5-turbo gpt-4 claude-3-sonnet; do
  echo "Testing with $model"
  aia --model $model code_review test_file.py
done

# Test parameter variations
aia code_review --file test1.py --severity high
aia code_review --file test2.py --severity low

Best Practices

Prompt Structure

  1. Clear Instructions: Specific, actionable directions
  2. Context Setting: Provide necessary background
  3. Output Format: Specify desired response structure
  4. Examples: Include sample inputs/outputs when helpful
  5. Error Handling: Account for edge cases

Configuration Management

  1. Model Selection: Choose appropriate models for tasks
  2. Temperature Setting: Adjust creativity vs. consistency
  3. Token Limits: Balance completeness with cost
  4. Parameter Validation: Ensure required inputs are provided

Maintenance

  1. Version Control: Track prompt changes
  2. Documentation: Keep usage instructions current
  3. Performance Monitoring: Track effectiveness over time
  4. User Feedback: Incorporate user suggestions

Explore the specific categories to find prompts that match your needs, or use these as inspiration to create your own custom prompts!