Breaking the Model Context Protocol: Agentic Attacks and Defenses for McP-Powered AI Systems
暫譯: 打破模型上下文協議：針對 McP 驅動的 AI 系統的代理攻擊與防禦

Satheesh Kumar, Thejes Sree, Sekar, Srinivasan

Breaking the Model Context Protocol: Agentic Attacks and Defenses for McP-Powered AI Systems
  • 出版商: Apress
  • 出版日期: 2026-08-26
  • 售價: $1,530
  • 貴賓價: 9.5 折 $1,453
  • 語言: 英文
  • 頁數: 277
  • 裝訂: Quality Paper - also called trade paper
  • ISBN: 9798868829673
  • ISBN-13: 9798868829673
  • 相關分類: Penetration-test

    • 海外代購書籍(需單獨結帳)

相關主題

商品描述

As AI agents plug into more tools and internal systems, the Model Context Protocol (MCP) is becoming a core part of how modern platforms work. With this shift comes a fast-growing challenge: understanding the new attack surfaces created when probabilistic models interact with real APIs, data, and networks. This book gives practitioners a clear, practical guide to navigating that emerging threat landscape by showing how MCP architectures behave in production and where hidden risks often appear.

The book begins by mapping today's MCP trust boundaries and explaining why traditional security assumptions don't hold when the "client" is an LLM. You'll explore real attack stories and hands-on labs demonstrating tool-poisoning techniques, signature cloaking, and sampling-based abuses. You'll then learn how attackers target the surrounding environment through DNS rebinding, malicious MCP servers, and confused-deputy patterns that turn over-permissioned tools into high-impact attack paths.

From there, the book provides defensive approaches built on schemas, contracts, monitoring, least privilege, and continuous red-team testing. Each chapter helps you apply the ideas to real deployments. Drawing on active MCP security research and real-world agent testing, this book offers a focused roadmap for securing the next generation of AI systems.

What You Will Learn

  • Understand how MCP architectures function in real AI agent systems
  • Identify trust boundaries and map emerging attack surfaces
  • Use sampling-based and elicitation-based techniques to assess model behavior
  • Protect MCP environments from DNS rebinding and confused-deputy risks

Who This Book is For

This book is for security engineers, AI platform teams, red-teamers, DevSecOps practitioners, MCP implementers, agent-framework developers, and technical leaders responsible for securing AI-driven systems and LLM-powered applications.

商品描述(中文翻譯)

隨著 AI 代理程式連接到更多工具和內部系統，模型上下文協議（Model Context Protocol, MCP）正成為現代平台運作的核心部分。隨著這一轉變，出現了一個快速增長的挑戰：理解當概率模型與真實的 API、數據和網絡互動時所產生的新攻擊面。本書為實務工作者提供了一個清晰、實用的指南，幫助他們在這個新興的威脅環境中導航，展示 MCP 架構在生產環境中的行為以及隱藏風險常出現的地方。

本書首先繪製當前 MCP 的信任邊界，並解釋為何當「客戶端」是大型語言模型（LLM）時，傳統的安全假設不再成立。您將探索真實的攻擊故事和實作實驗室，展示工具中毒技術、簽名隱蔽和基於取樣的濫用。接著，您將學習攻擊者如何通過 DNS 重新綁定、惡意 MCP 伺服器和混淆代理模式來針對周圍環境，將過度授權的工具轉變為高影響力的攻擊路徑。

接下來，本書提供了基於架構、合約、監控、最小權限和持續紅隊測試的防禦方法。每一章都幫助您將這些理念應用於實際部署。基於活躍的 MCP 安全研究和真實世界的代理測試，本書提供了一個專注的路線圖，以保護下一代 AI 系統的安全。

**您將學到什麼**

- 了解 MCP 架構在真實 AI 代理系統中的運作方式
- 確定信任邊界並繪製新興攻擊面
- 使用基於取樣和引導的技術來評估模型行為
- 保護 MCP 環境免受 DNS 重新綁定和混淆代理風險

**本書適合誰**

本書適合安全工程師、AI 平台團隊、紅隊成員、DevSecOps 實務者、MCP 實施者、代理框架開發者以及負責保護 AI 驅動系統和 LLM 驅動應用程式的技術領導者。

作者簡介

Thejes sree Satheesh kumar is a Quality Analyst - Consultant at ThoughtWorks, specialising in application and AI security testing. She is a Certified Ethical Hacker and holds CompTIA Security+, ISC2 Certified in Cybersecurity and Google Cybersecurity Professional certifications. With a strong background in automation testing using Playwright, Selenium, WebdriverIO, and Appium, Thejes combines quality engineering and security practices to build resilient software systems. She is passionate regarding secure AI ecosystems and advancing defensive strategies for emerging technologies like the Model Context Protocol (MCP). She is a speaker at various conferences, including NullCon and TechXpresso.

Srinivasan Sekar is an AI enthusiast and the Director of Engineering at TestMu AI (formerly LambdaTest), where he leads innovation in Agentic AI. His work focuses on building next-generation AI platforms and leveraging the Model Context Protocol (MCP) to create intelligent agentic applications. A passionate advocate for open source, Srinivasan is a recognised Appium member and an active contributor to several prominent projects, including Selenium, Appium, and Webdriver.io. He is a frequent speaker at international technology conferences, providing his deep expertise at events such as SeleniumConf, AppiumConf, and FOSDEM on the architecture and practical application of emerging AI technologies

作者簡介(中文翻譯)

Thejes sree Satheesh kumar 是 ThoughtWorks 的品質分析師 - 顧問，專注於應用程式和人工智慧安全測試。她是認證的道德駭客，並持有 CompTIA Security+、ISC2 網路安全認證以及 Google 網路安全專業認證。Thejes 擁有使用 Playwright、Selenium、WebdriverIO 和 Appium 進行自動化測試的堅實背景，結合品質工程和安全實踐來構建韌性軟體系統。她對安全的人工智慧生態系統充滿熱情，並致力於推進新興技術的防禦策略，如模型上下文協議 (Model Context Protocol, MCP)。她是多個會議的演講者，包括 NullCon 和 TechXpresso。

Srinivasan Sekar 是一位人工智慧愛好者，擔任 TestMu AI（前身為 LambdaTest）的工程總監，負責引領 Agentic AI 的創新。他的工作專注於構建下一代人工智慧平台，並利用模型上下文協議 (Model Context Protocol, MCP) 來創建智能的代理應用程式。作為開源的熱情倡導者，Srinivasan 是一位公認的 Appium 成員，並積極參與多個知名專案，包括 Selenium、Appium 和 Webdriver.io。他經常在國際技術會議上發表演講，分享他在新興人工智慧技術架構和實際應用方面的深厚專業知識，參加的活動包括 SeleniumConf、AppiumConf 和 FOSDEM。

類似商品

最後瀏覽商品 (20)