现在 AI Agent 概念满天飞，可真要落地复杂业务，一堆硬骨头就全冒出来了——架构怎么搭、记忆怎么管、多 Agent 怎么协调才不互相拆台。

越写越乱，最后代码直接糊成一锅粥，上线更是想都别想。

看到一本开源书，叫 Agentic Design Patterns，专治这种无从下手的迷茫。

这套体系相当不错： ① 21 章正文加 7 个附录，按难度拆成四个部分，从入门到企业级全链路覆盖 ② 每章都配了 Jupyter Notebook，不是让你干瞪眼，而是边读边跑代码，理论和实操焊死在一起 ③ 前半程打地基：提示链、路由、并行、反思、工具调用、多 Agent 协作，核心模式一次讲透 ④ 后半程直接对准生产环境：记忆管理、异常恢复、人机协作、安全护栏、性能评估，上线前可能踩的坑基本都替你蹚平了 ⑤ 附录里还塞了框架对比和高级提示技巧，随手翻翻就能补漏 从“会用 Agent”到“能设计 Agent 系统”，差的就是这套系统方法论。

https://github.com/evoiz/Agentic-Design-Patterns…

evoiz/Agentic-Design-Patterns

Source: https://github.com/evoiz/Agentic-Design-Patterns

📚 Agentic Design Patterns - A Hands-On Guide to Building Intelligent Systems

📖 About This Repository

This repository contains the complete materials for “Agentic Design Patterns: A Hands-On Guide to Building Intelligent Systems” by Antonio Gulli. It includes all chapters in PDF format and accompanying code notebooks for hands-on learning.

Note: All author royalties are donated to Save the Children 💝

🎯 What You’ll Learn

This comprehensive guide covers 21 chapters and 7 appendices on building intelligent AI agent systems, including:

• Foundational Patterns: Prompt chaining, routing, parallelization
• Advanced Techniques: Reflection, tool use, planning, multi-agent systems
• Memory & Learning: Memory management, adaptation, goal setting
• Production Patterns: Exception handling, human-in-the-loop, RAG
• Optimization: Resource-aware patterns, reasoning techniques, guardrails
• Real-world Applications: From GUI to real-world environments

📁 Repository Structure

.
├── 📄 README.md                           # This file
├── 📚 book/
│   └── Agentic_Design_Patterns_Complete.pdf  # Complete book (424 pages)
├── 💻 chapter_notebooks/                          # Chapter 
│   ├── Chapter_01_Prompt_Chaining.ipynb
│   ├── Chapter_02_Routing.ipynb
│   ├── Chapter_03_Parallelization.ipynb
│   ├── ...
│   └── Appendix_G_Coding_Agents.ipynb

📚 Table of Contents

Introduction & Foundations

• Dedication
• Acknowledgment
• Foreword
• A Thought Leader’s Perspective: Power and Responsibility
• Introduction
• What makes an AI system an “agent”?

Part One: Core Patterns (103 pages)

1. Chapter 1: Prompt Chaining - Sequential task decomposition
2. Chapter 2: Routing - Dynamic path selection
3. Chapter 3: Parallelization - Concurrent processing
4. Chapter 4: Reflection - Self-improvement mechanisms
5. Chapter 5: Tool Use - External capability integration
6. Chapter 6: Planning - Strategic task management
7. Chapter 7: Multi-Agent - Collaborative systems

Part Two: Advanced Patterns (61 pages)

8. Chapter 8: Memory Management - State persistence
9. Chapter 9: Learning and Adaptation - Dynamic improvement
10. Chapter 10: Model Context Protocol (MCP) - Standardized interfaces
11. Chapter 11: Goal Setting and Monitoring - Objective tracking

Part Three: Production Patterns (34 pages)

12. Chapter 12: Exception Handling and Recovery - Robust error management
13. Chapter 13: Human-in-the-Loop - Human-AI collaboration
14. Chapter 14: Knowledge Retrieval (RAG) - Information access patterns

Part Four: Enterprise Patterns (114 pages)

15. Chapter 15: Inter-Agent Communication (A2A) - Agent networking
16. Chapter 16: Resource-Aware Optimization - Efficient resource usage
17. Chapter 17: Reasoning Techniques - Advanced decision-making
18. Chapter 18: Guardrails/Safety Patterns - Risk mitigation
19. Chapter 19: Evaluation and Monitoring - Performance tracking
20. Chapter 20: Prioritization - Task management
21. Chapter 21: Exploration and Discovery - Autonomous learning

Appendices (74 pages)

• Appendix A: Advanced Prompting Techniques
• Appendix B: AI Agentic: From GUI to Real world environment
• Appendix C: Quick overview of Agentic Frameworks
• Appendix D: Building an Agent with AgentSpace
• Appendix E: AI Agents on the CLI
• Appendix F: Under the Hood: Reasoning Engines
• Appendix G: Coding agents

Conclusion & References

• Conclusion
• Glossary
• Index of Terms

🚀 Getting Started

Prerequisites

# Python 3.8 or higher required
python --version

# Install Jupyter for notebooks
pip install jupyter notebook

# Install common dependencies
pip install -r requirements.txt

Installation

1. Clone the repository

git clone https://github.com/evoiz/Agentic-Design-Patterns.git
cd Agentic-Design-Patterns.git

2. Set up virtual environment (recommended)

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

pip install jupyter notebook
pip install pandas numpy matplotlib openai langchain

4. Launch Jupyter Notebook

jupyter notebook

💻 Running the Code

Each chapter includes a Jupyter notebook with practical examples:

1. Navigate to the chapter_notebooks/ directory
2. Open the desired chapter notebook
3. Follow the instructions within each notebook
4. Run cells sequentially for best learning experience

Example: Running Chapter 1

# Navigate to notebooks directory
cd chapter_notebooks

# Launch specific notebook
jupyter notebook Chapter_01_Prompt_Chaining.ipynb

📖 How to Use This Repository

For Self-Study

1. Read each chapter in the PDF
2. Open the corresponding notebook
3. Run the code examples
4. Experiment with modifications
5. Complete the exercises

For Teaching

1. Use chapters as lecture materials
2. Assign notebooks as lab exercises
3. Create custom examples based on patterns
4. Build projects using multiple patterns

For Research

1. Reference implementation patterns
2. Benchmark different approaches
3. Extend patterns for new use cases
4. Contribute improvements back

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines.

Ways to Contribute

• 🐛 Report bugs and issues
• 💡 Suggest new features or patterns
• 📝 Improve documentation
• 🔧 Submit code improvements
• 🌍 Translate materials

Contribution Process

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📚 Additional Resources

Official Links

• 📖 Pre-order the Book
• 👨‍💼 Author’s LinkedIn
• 📁 Original Google Drive Materials

Related Frameworks

• LangChain
• AutoGPT
• OpenAI Assistants
• Microsoft AutoGen
• CrewAI

Learning Path

1. Beginners: Start with Chapters 1-7 (Core Patterns)
2. Intermediate: Progress through Chapters 8-14 (Advanced & Production)
3. Advanced: Master Chapters 15-21 (Enterprise Patterns)
4. Experts: Explore Appendices for cutting-edge techniques

⚖️ License

This repository is for educational purposes. Please respect the author’s copyright and intellectual property rights.

• Book Content: © Antonio Gulli - All rights reserved
• Code Examples: MIT License (see LICENSE file)
• Educational Use: Permitted with attribution

🙏 Acknowledgments

• Antonio Gulli - Author and AI thought leader
• Save the Children - Beneficiary of all book royalties
• Contributors - Everyone who helps improve these materials
• Community - Learners and practitioners advancing AI agents

📞 Contact & Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Author: LinkedIn

🌟 Star History

If you find this repository useful, please consider giving it a star ⭐

📊 Repository Stats

Building the future of AI, one pattern at a time 🚀

Made with ❤️ for the AI community