Description
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 fastâgrowing 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 handsâon labs demonstrating toolâpoisoning techniques, signature cloaking, and samplingâbased abuses. You’ll then learn how attackers target the surrounding environment through DNS rebinding, malicious MCP servers, and confusedâdeputy patterns that turn overâpermissioned tools into highâimpact attack paths.
From there, the book provides defensive approaches built on schemas, contracts, monitoring, least privilege, and continuous redâteam testing. Each chapter helps you apply the ideas to real deployments. Drawing on active MCP security research and realâworld 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 samplingâbased and elicitationâbased techniques to assess model behavior
- Protect MCP environments from DNS rebinding and confusedâdeputy risks
Who This Book is For
This book is for security engineers, AI platform teams, redâteamers, DevSecOps practitioners, MCP implementers, agentâframework developers, and technical leaders responsible for securing AIâdriven systems and LLMâpowered applications.
Table of Contents
Part I – Foundations: The Agentic Threat Landscape.- Chapter 1: From Architecture to Attack Surface.- Chapter 2: Deconstructing the MCP Trust Boundaries.- Chapter 3: Thinking Like an Attacker (The Red Team Mindset).- Part II – Tool and MCP-Specific Abuse.- Chapter 4: Advanced Tool Poisoning Attack (ATPA) – Output Poisoning.- Chapter 5: Signature Cloaking: The Invisible MCP Parameters.- Part III – Sampling and Elicitation Attacks.- Chapter 6: Sampling Abuse and the Hijacked Assistant.- Chapter 7: Elicitation Attacks: Weaponising Forms and URLs.- Part IV – Network and Environment Abuse.- Chapter 8: DNS Rebinding and Turning Your Agent Against Your Own Network.- Chapter 9: The Confused Deputy: The Agent With Too Many Keys.- Part V – WebMCP.- Chapter 10: WebMCP and the Browser as an MCP Server.- Chapter 11: WebMCP Security and the Authenticated Attack Surface.- Part VI – Putting It All Together.- Chapter 12: The Defender’s Pocket Playbook.



