Agentic AI for Cybersecurity: Foundations, Frameworks, and Field Practice
暫譯: 網路安全 Agentic AI：基礎、框架與實務應用
Singh, Amit
- 出版商: Apress
- 出版日期: 2026-10-23
- 售價: $2,180
- 貴賓價: 9.5 折 $2,071
- 語言: 英文
- 頁數: 562
- 裝訂: Quality Paper - also called trade paper
- ISBN: 9798868832369
- ISBN-13: 9798868832369
-
相關分類:
Penetration-test、Large language model、Computer-networks
海外代購書籍(需單獨結帳)
商品描述
Agentic AI is no longer theoretical--but most cybersecurity teams are still struggling to move from concept to production. Agentic AI for Cybersecurity fills the critical gap between vendor hype and academic abstraction, delivering a practical, engineering-first guide that security architects can apply immediately. This is the book for professionals who need to build real systems--not just talk about them.
At its core, this book argues that agentic AI is not a toolset but a new operating model for the Security Operations Center. That shift changes everything: architecture, governance, and risk. Unlike other titles, it refuses to separate capability from security. Offensive and defensive use cases are treated together, reflecting the real dual-use nature of agentic systems and equipping readers to design--and defend--against both.
Timeliness is what makes this book essential. It is the first practitioner-focused guide to unify three converging forces shaping modern security: MCP-based agentic systems, their emerging and largely undocumented attack surface, and the parallel migration to post-quantum cryptography. Each is complex on its own--together, they define the next generation of security architecture. This book addresses them as a single, integrated challenge.
Built for real-world use, every chapter follows a 'what / how / why' structure and culminates in runnable, production-grade architectures. Readers don't just learn concepts--they deploy them, using reference implementations drawn from actual environments and supported by a living GitHub repository. From MCP servers and adversary emulation to AI-driven detection and quantum-resistant infrastructure, this is hands-on guidance grounded in reality.
For practitioners who need clarity, credibility, and actionable design patterns, Agentic AI for Cybersecurity stands apart. It is unapologetically practical, rigorously honest about risks, and singular in its integration of architecture, security, and emerging AI paradigms--making it a must-have resource for building the next generation of secure systems.
What You Will Learn:
- Master the conceptual architecture of agentic AI
- Build, harden, and govern production-grade Model Context Protocol (MCP) servers for network and security operations
- Operationalize AI-driven offensive security through Kali Linux integration via MCP and MITRE Caldera as an LLM-driven adversary emulation platform
- Defend against the agentic AI attack surface
- Plan and execute the post-quantum cryptographic migration in AI-era infrastructure
- Apply machine learning and deep learning to autonomous malware detection and threat classification
Who this Book is for:
The target reader is a working network or security practitioner -- a SOC analyst, network engineer, security architect, security engineer, or incident responder -- who is being asked, by their employer or by the field, to integrate agentic AI into their operational practice and who needs guidance that is more concrete than vendor whitepapers and more applicable than academic papers.
The book does not assume prior experience with agentic AI specifically, but it does assume the reader is comfortable with at least one of the three domains the book unifies: enterprise network operations, security operations, or applied machine learning. Readers from adjacent communities -- DevSecOps engineers, cloud security architects, AI/ML platform engineers responsible for AI safety -- will also find Part II directly applicable to their work.
商品描述(中文翻譯)
Agentic AI 已不再只是理論，但多數資安團隊仍在努力將概念落實到正式環境。《Agentic AI for Cybersecurity》填補了供應商炒作與學術抽象之間的重要缺口，提供一套以實務與工程為核心的指南，讓資安架構師能立即應用。本書適合需要建構真正系統，而不只是討論系統的專業人士。
本書的核心論點是：Agentic AI 並不是一組工具，而是 Security Operations Center 的全新運作模式。這項轉變將影響一切，包括架構、治理與風險。不同於其他書籍，本書拒絕將功能與安全性分開討論。書中將攻擊性與防禦性使用情境一併探討，反映 Agentic 系統真實的雙重用途特性，並協助讀者同時設計系統，以及防禦這類系統所帶來的威脅。
本書的重要性來自其時效性。這是第一本以實務工作者為核心、整合三股正在匯聚且形塑現代資安的力量的指南：以 MCP 為基礎的 Agentic 系統、其正在形成且大多尚未有完整文件記載的攻擊面，以及同步進行中的後量子密碼學遷移。這三者各自都十分複雜，而合在一起，便定義了下一世代的資安架構。本書將它們視為單一且整合性的挑戰加以處理。
本書以真實世界的應用為導向，每一章都採用「是什麼／如何做／為什麼」的結構，最後以可執行、符合正式環境需求的架構作結。讀者不只是學習概念，更能實際部署這些概念；書中使用來自真實環境的參考實作，並由持續更新的 GitHub repository 提供支援。從 MCP servers 與 adversary emulation，到 AI 驅動的偵測機制與抗量子基礎架構，本書提供的是立足現實的實作導向指南。
對於需要清晰度、可信度與可實際採用之設計模式的實務工作者而言，《Agentic AI for Cybersecurity》具有鮮明的優勢。本書毫不掩飾地強調實務應用，並以嚴謹且誠實的態度面對風險；同時，它獨特地整合了架構、安全性與新興 AI 範式，是建構下一世代安全系統不可或缺的資源。
您將學到：
• 掌握 Agentic AI 的概念架構
• 建構、強化與治理適用於正式環境的 Model Context Protocol（MCP）servers，用於網路與資安作業
• 透過 MCP 整合 Kali Linux，並運用 MITRE Caldera 作為由 LLM 驅動的 adversary emulation platform，將 AI 驅動的攻擊性資安作業導入實際運作
• 防禦 Agentic AI 的攻擊面
• 規劃並執行 AI 時代基礎架構中的後量子密碼學遷移
• 將 machine learning 與 deep learning 應用於自主式 malware detection 與威脅分類
本書適合哪些讀者：
本書的目標讀者是在職的網路或資安從業人員，例如 SOC analyst、network engineer、security architect、security engineer 或 incident responder。他們可能正受到雇主或整個產業的要求，需要將 Agentic AI 整合至日常作業實務中，並需要比供應商白皮書更具體、比學術論文更能直接應用的指引。
本書不假設讀者具備 Agentic AI 的先前經驗，但假設讀者至少熟悉本書所整合之三個領域中的一項：企業網路作業、資安作業或應用機器學習。來自相鄰領域的讀者，例如 DevSecOps engineer、cloud security architect，以及負責 AI 安全性的 AI/ML platform engineer，也會發現第二部的內容能直接應用於其工作。
作者簡介
Amit Singh is a Lead Architect for Secure Networking and a triple CCIE with more than 20 years of experience designing, securing, and scaling complex network infrastructures across enterprise networking, cybersecurity, and software-defined architectures.
His current work sits at the intersection of agentic AI and cybersecurity -- the engineering, hardening, and governance of Model Context Protocol (MCP) servers as the integration layer for AI-driven security operations. Among the ways he contributes to the open-source community is building production MCP servers for security workflows, several of which are published in the Cisco DevNet exchange. His approach pairs protocol-level depth with production-ready patterns -- turning what is usually vendor whitepaper material into infrastructure that actually ships into a security operations center. His original agentic-orchestration framework, which underpins this book, is published as a defensive technical disclosure (Technical Disclosure Commons, 2026).
Under the SDN Tech Forum brand -- a YouTube channel, a Spotify podcast, and a Medium publication -- Amit's educational content has been viewed more than a million times by networking professionals worldwide. He has delivered keynotes and mentored engineers on Zero-Trust, SDN, cloud integration, and, most recently, agentic AI architectures and the security posture they require. He holds a Master of Science in Computer Networks from North Carolina State University.
Agentic AI for Cybersecurity is written from a practitioner's perspective -- grounded in architecture and validated by deployment. Its chapters on MCP server construction, Kali Linux integration, MITRE Caldera adversary emulation, AI-powered malware detection, and post-quantum cryptography migration are drawn directly from Amit's field work.
LinkedIn: linkedin.com/in/amit-singh-5b774754
YouTube: @SDNTechForum
Amazon: amazon.com/stores/author/B0GNHTW3CV
作者簡介(中文翻譯)
Amit Singh 是 Secure Networking 領域的首席架構師，同時擁有三項 CCIE 認證，具備超過 20 年的經驗，專長於企業網路、資安與軟體定義架構，負責設計、保護及擴展複雜的網路基礎架構。
他目前的工作位於代理式 AI（agentic AI）與資安的交會點，專注於 Model Context Protocol（MCP）伺服器的工程實作、強化與治理，將其打造為 AI 驅動安全作業的整合層。他對開放原始碼社群的貢獻之一，是為資安工作流程建置可用於正式環境的 MCP 伺服器，其中數個已發布於 Cisco DevNet exchange。他的方法結合了協定層級的深度理解與可供正式環境採用的設計模式，將通常只存在於廠商白皮書中的內容，轉化為真正能部署至安全作業中心的基礎架構。他所建立、並作為本書基礎的原創代理式協調框架，已以防禦性技術揭露的形式發布於 Technical Disclosure Commons（2026）。
在 SDN Tech Forum 品牌旗下——包括 YouTube 頻道、Spotify podcast 與 Medium 出版平台——Amit 製作的教育內容已獲得全球網路專業人士超過百萬次觀看。他曾針對 Zero-Trust、SDN、雲端整合，以及近期的代理式 AI 架構與其所需的安全態勢，發表主題演講並指導工程師。他擁有 North Carolina State University 的 Computer Networks 理學碩士學位。
《Agentic AI for Cybersecurity》以實務工作者的觀點撰寫，內容以架構為基礎，並透過實際部署加以驗證。本書關於 MCP 伺服器建置、Kali Linux 整合、MITRE Caldera 對手模擬、AI 驅動的惡意軟體偵測，以及後量子密碼學遷移的章節，皆直接取材自 Amit 的實務經驗。
LinkedIn：linkedin.com/in/amit-singh-5b774754
YouTube：@SDNTechForum
Amazon：amazon.com/stores/author/B0GNHTW3CV
