Volume 2 of The Agentic AI Handbook. This comprehensive guide equips AI architects, software engineers, and technical leaders with the blueprints needed to design, govern, and scale autonomous systems. Moving beyond basic prompts, this volume dives deep into advanced architectural patterns like dynamic planning, goal decomposition, and critic-refiner loops. Readers will explore the mechanics of multi-agent collaboration, mastering role design, delegation, and communication protocols while avoiding common coordination anti-patterns.
As agents interact with the real world, robust governance becomes critical. This book provides practical strategies for implementing computer-use and browser automation agents, balancing GUI-versus-API tradeoffs, and grounding screen understanding. You will learn how to secure these systems through sandboxing, risk policy gates, and human-in-the-loop approval checkpoints. Furthermore, it covers rigorous evaluation methodologies, guiding you through hallucination containment, telemetry, and industry-standard agent benchmarking using SWE-bench, GAIA, and OSWorld to ensure cost-adjusted accuracy and production observability.
Finally, the text navigates the rapidly evolving landscape of infrastructure and protocols. It offers a detailed breakdown of leading frameworks, including LangGraph, CrewAI, Microsoft AutoGen, and the latest SDKs from OpenAI, Claude, and Google. By mastering emerging standards like the Model Context Protocol, Agent2Agent communication, and Agent Payments, technical professionals will be fully prepared to choose the right interoperability stack and build the next generation of secure, collaborative AI systems.
外文書商品之書封，為出版社提供之樣本。實際出貨商品，以出版社所提供之現有版本為主。部份書籍，因出版社供應狀況特殊，匯率將依實際狀況做調整。
無庫存之商品，在您完成訂單程序之後，將以空運的方式為你下單調貨。為了縮短等待的時間，建議您將外文書與其他商品分開下單，以獲得最快的取貨速度，平均調貨時間為1~2個月。
為了保護您的權益，「三民網路書店」提供會員七日商品鑑賞期(收到商品為起始日)。
若要辦理退貨，請在商品鑑賞期內寄回，且商品必須是全新狀態與完整包裝(商品、附件、發票、隨貨贈品等)否則恕不接受退貨。
