A comprehensive collection of AI prompt engineering resources: distilled best-practice guides, a professional prompt-engineering agent, and ready-to-use prompt examples.

🎯 Perfect for: Developers, AI practitioners, and anyone looking to improve their prompt engineering skills with proven techniques from industry leaders.

• 📚 Curated Guides — Distilled best practices from OpenAI, Anthropic, Google, and more
• 🤖 AI Agent — Transform simple requirements into production-ready prompts
• 📝 Ready-Made Examples — Copy-paste prompts for common use cases
• 🔧 Meta-Prompts — Advanced tools for creating and refining prompts

New to prompt engineering? Start here:

# 1. Clone the repository
git clone https://github.com/thibaultyou/prompt-blueprint.git
cd prompt-blueprint

# 2. Read the unified best practices guide (recommended first read)
open guides/unified-best-practices__claude_sonnet_4.md

# 3. Try the prompt engineering agent
open meta-prompts/prompt-engineering-agent.md
# Copy the content and paste it into ChatGPT, Claude, or your preferred AI

# 4. Explore ready-made examples
open examples/customer-support-agent.md

Already experienced? Jump to the prompt engineering agent or browse specific guides for your AI model.

What are these? Concise, actionable guides that distill official documentation and best practices from major AI providers into practical advice you can use immediately.

1. 🌟 Start with the unified guide — Get a comprehensive overview of all best practices
2. 🎯 Choose provider-specific guides — Dive deeper into techniques for your preferred AI model
3. 🔄 Reference during prompt creation — Use alongside our prompt-engineering agent
4. 📈 Iterate and improve — Combine insights from multiple guides for optimal results

Important: These are AI-generated summaries of official documentation, optimized for practical use. They are not official provider documents but are based on official sources.

What is it? A sophisticated AI agent that transforms your simple requirements into professional, production-ready prompts with documentation and usage guidelines.

flowchart LR
    A[👤 Your Request<br/>I need a customer service prompt] --> B{Add Context?}
    
    B -->|No| C[🤖 Agent Only]
    B -->|Yes| D[📚 Unified Guide<br/>+ 🤖 Agent]
    
    C --> E[✅ Good Prompt]
    D --> F[⭐ Better Prompt<br/>with advanced techniques]
    
    style A fill:#e3f2fd
    style D fill:#f3e5f5
    style F fill:#fff9c4

Simple workflow:

1. Ask the agent what you need
2. Optional: Load the unified guide first for enhanced results
3. Get your prompt — ready to use!

• ✅ Save hours — No need to research best practices or iterate manually
• ✅ Professional quality — Get enterprise-grade prompts with proper formatting
• ✅ Cross-platform — Works with GPT-4, Claude, Gemini, and other models
• ✅ Complete package — Includes usage guidelines, troubleshooting, and metrics

1. Open meta-prompts/prompt-engineering-agent.md
2. Copy the entire content
3. Paste it into your preferred AI interface (ChatGPT, Claude, etc.)

💡 Pro Tip: For even better results, first load the unified-best-practices__claude_sonnet_4.md guide to give the agent additional context about advanced techniques.

Just tell the agent what kind of prompt you want:

Simple examples:

• "I need a customer service chatbot prompt"
• "Create a code review prompt for my team"
• "Design a creative writing prompt for marketing copy"

Detailed examples (even better results):

• "I need a customer support prompt for a SaaS company, handling 100+ daily conversations, optimized for Claude"
• "Create a technical documentation prompt for API references that must be GDPR compliant"

The agent delivers a complete package:

• 📝 Production-ready prompt with professional formatting
• 📋 Usage guidelines and implementation notes
• 📊 Performance benchmarks and success metrics
• 🔧 Troubleshooting guide for common issues
• 🌐 Cross-platform compatibility notes

Ask for specific adjustments:

• "Make it more concise for high-volume use"
• "Add compliance requirements for healthcare"
• "Optimize specifically for GPT-4"

You: "I need a prompt for customer service chatbot that handles complaints professionally"

Agent: [Executes 7-stage professional workflow]
       [Delivers complete prompt package with documentation]

You: "Great! Can you make it more concise for high-volume use?"

Agent: [Provides optimized version focused on efficiency]

• prompt-engineering-agent.md — The main agent that creates professional prompts from your requirements
• documentation-expert-agent.md — The agent that creates and updates the guides themselves
• customer-support-agent.md — Example of a production-ready prompt created by the agent

1. 📖 Research & Distillation — The documentation-expert-agent processes official documentation from AI providers and creates the guides in /guides

2. 🧠 Knowledge Integration — The prompt-engineering-agent embodies these best practices, using them to create high-quality prompts

3. 📚 Example Generation — The agent's outputs are collected in /examples as ready-to-use prompts

4. 🔄 Continuous Improvement — As the agents evolve, the guides and examples are updated to reflect the latest best practices

We welcome contributions! Here's how you can help:

• 🐛 Report issues — Found a problem? Let us know
• 📝 Suggest improvements — Ideas for better guides or examples
• 🔧 Submit pull requests — New guides, examples, or improvements
• 💡 Share feedback — How are you using these resources?

Note: This is a personal research project that evolves over time. It may not reach production-grade maturity but aims to be a valuable learning resource.

• 🚧 Work in Progress — This repository is provided as-is for educational purposes
• 📄 Unofficial Content — Guides are AI-generated summaries, not official provider documentation
• 🔒 No Warranties — Use at your own discretion; no guarantees for production use
• 🎓 Educational Focus — Designed for learning and experimentation, not commercial deployment

• Check the guides for fundamentals
• Try the prompt engineering agent for custom prompts
• Browse examples for inspiration
• Open an issue for specific problems or suggestions