AI Agent Architecture - The Book

From Concept to Production: Framework-Agnostic AI Agent Architecture Patterns

Author: Wayland Zhang

---

Languages / 语言 / 言語

Language	Status	Link
中文 (Chinese)	Complete	阅读中文版
日本語 (Japanese)	Complete	日本語版を読む
English	Complete	Read English Version

---

About This Book

This book is a practical guide to understanding AI Agent system design patterns, not just another framework tutorial.

9 Parts, 33 Chapters covering:

Part	Topic
Part 1	Agent Fundamentals (ReAct Loop)
Part 2	Tools & Extensions (MCP, Skills, Hooks)
Part 3	Context & Memory
Part 4	Single Agent Patterns (Planning, Reflection, CoT)
Part 5	Multi-Agent Orchestration (DAG, Supervisor, Handoff)
Part 6	Advanced Reasoning (ToT, Debate, Research)
Part 7	Production Architecture
Part 8	Enterprise Features (Token Budget, OPA, WASI Sandbox)
Part 9	Frontier Practices (Computer Use, Agentic Coding)

---

Reference Implementation

This book uses Shannon as a reference implementation for multi-agent orchestration, and ShanClaw as the local Agent Harness reference (→ Chapter 33).

Shannon is a three-layer multi-agent system:

Orchestrator (Go)    - Orchestration, Budget, Policy
Agent Core (Rust)    - Execution, Sandbox, Rate Limiting
LLM Service (Python) - Inference, Tools, Vectors

---

Citation

If you use this book in your research or work, please cite it:

@book{zhang2026aiagent,
  author = {Zhang, Wayland},
  title = {AI Agent Architecture: From Monolith to Enterprise Multi-Agent},
  year = {2026},
  url = {https://waylandz.com/ai-agent-book},
  publisher = {Self-published}
}

Or in APA format:

Zhang, Wayland. (2026). AI Agent Architecture: From Monolith to Enterprise Multi-Agent. https://waylandz.com/ai-agent-book

---

License

Book content: CC BY-NC-SA 4.0

Shannon OSS code: Apache 2.0