Agentic AI for Cybersecurity: Building Autonomous Defenders and Adversaries
Unlock the power of next‑generation AI agents to transform cybersecurity, business operations, and productivity.
Agentic AI has transformed cybersecurity and other domains with autonomous, intelligent software agents that can plan, reason, and take action. Instead of simply answering questions, these agents work like digital analysts using tools and adapting to new information, while collaborating with humans and other agents to get real work done.
Agentic AI for Cybersecurity is a practical, hands‑on guide to building, deploying, and securing these systems. Written in plain English, the book explains how modern agentic AI frameworks (such as LangChain, LangGraph, CrewAI, and LlamaIndex) enable autonomous, multi‑agent systems, and how frameworks like Model Context Protocol (MCP) and agent‑to‑agent (A2A) are creating an emerging Internet of Agents.
Through step‑by‑step projects and real‑world case studies, readers learn how to apply agentic AI to cybersecurity defense, adversarial testing, business automation, and productivity workflows. The book also addresses critical risks, including adversarial threats, governance, and compliance. The goal is to make security a first‑class concern, not an afterthought.
Designed for both technical and non‑technical readers, this book bridges the gap between AI theory and real‑world impact.
Key Features
- Hands‑on projects with practical examples for building and deploying AI agents
- Covers leading agentic AI frameworks and protocols (LangChain, CrewAI, LangGraph, MCP, A2A, n8n, OpenAI Agent Builder, and more)
- Security‑first approach with dedicated coverage of adversarial threats and governance
- Plain‑English explanations with diagrams, exercises, and GitHub resources
Who This Book Is For
Cybersecurity professionals, AI engineers, developers, architects, business and technology leaders, or anyone looking to understand and apply agentic AI responsibly and effectively.
Table of Contents:
Introduction.. . . . . . . . . . . . . . . xvii
Part I Foundations and Core Technologies
Chapter 1 Foundations of Agentic AI Systems. . . . . . . . . 1
Reflecting and Defining the Shift.. . . . . . . . . . 2
The Anatomy of a Modern AI Agentic Infrastructure.. . . . . . 9
Summary. . . . . . . . . . . . . . 13
Review Questions.. . . . . . . . . . . . . 15
Answers to Review Questions. . . . . . . . . . 17
Chapter 2 Retrieval-Augmented Generation (RAG), Agentic RAG, and Model Context Protocol (MCP).. . . . . . . . . . 19
A Recap of Retrieval-Augmented Generation (RAG) and Agentic RAG. . . 20
Vector Databases and Embedding Models. . . . . . . . 21
A Deep Dive into the Model Context Protocol (MCP).. . . . . . 42
Summary. . . . . . . . . . . . . . 46
Review Questions.. . . . . . . . . . . . . 47
Answers to Review Questions. . . . . . . . . . 50
End-of-Chapter Projects.. . . . . . . . . . . 51
Chapter 3 Protocols for the Internet of Agents.. . . . . . . . . 55
The Agent2Agent (A2A) Protocol . . . . . . . . . . 55
The Agent Communication Protocol (ACP). . . . . . . . 64
Agent Name Service (ANS): A Secure Universal Directory. . . . . 66
AGNTCY: Building the Foundational Stack for an “Internet of Agents”.. . . 69
Summary. . . . . . . . . . . . . . 71
Review Questions.. . . . . . . . . . . . . 72
Answers to Review Questions. . . . . . . . . . 74
Chapter 4 Orchestration with LangChain, LangGraph, and LlamaIndex. . . 75
Introducing LangChain. . . . . . . . . . . . 76
Using LangGraph Agent Workflows.. . . . . . . . . 78
Introducing LlamaIndex.. . . . . . . . . . . 87
Choosing Your Framework.. . . . . . . . . . . 89
Summary. . . . . . . . . . . . . . 91
Review Questions.. . . . . . . . . . . . . 92
Answers to Review Questions. . . . . . . . . . 94
Chapter 5 Exploring CrewAI, n8n, Apache Airflow, OpenAI Agent Builder, and Other Agentic Frameworks.. . . . . . . . . . 95
CrewAI: The Power of Role-Based Collaboration. . . . . . . 96
n8n: Visual Workflow and Low-Code Automation. . . . . . 101
Apache Airflow.. . . . . . . . . . . . . 116
Langflow: Visual Orchestration and Deployment for GenAI.. . . . 121
The OpenAI Agent Platform.. . . . . . . . . . 122
Summary. . . . . . . . . . . . . . 125
Chapter Review Questions.. . . . . . . . . . . 126
Answers to Review Questions. . . . . . . . . . 128
Part II Applications and Security
Chapter 6 The Architect’s Cockpit: AI-Powered IDEs and Coding Agents.. . . 131
Defining the Modern AI Coding Agent. . . . . . . . 131
A Comparative Analysis of Leading AI Coding Agents.. . . . . 139
Establishing Enterprise-Grade Evaluation Criteria. . . . . . 154
Project CodeGuard: Don’t Let Your AI Agents Introduce Security Vulnerabilities.. 156
Summary. . . . . . . . . . . . . . 158
Chapter Review Questions.. . . . . . . . . . . 159
Answers to Review Questions. . . . . . . . . . 161
Project 6-1: Using Project CodeGuard to Create More Secure Code. . . 162
Chapter 7 Building an Autonomous SOC Analyst. . . . . . . . 165
The Crisis in Modern Security Operations. . . . . . . . 166
Anatomy of an Autonomous SOC Analyst.. . . . . . . . 167
Building the Foundation: Technical Implementation.. . . . . . 171
Advanced Implementation Patterns. . . . . . . . . 197
Summary. . . . . . . . . . . . . . 210
Chapter Review Questions.. . . . . . . . . . . 211
Answers to Review Questions. . . . . . . . . . 212
Project 7-1: Building Your Own Autonomous SOC Analyst. . . . . 213
Chapter 8 Agentic Penetration Testing, Red Teaming, and Bug Bounties. . . 215
Real-Life Attacks.. . . . . . . . . . . . . 216
Case Study: Using LangGraph and n8n for Cyberattacks and Espionage.. . 218
Building Penetration Testing Agents with LangChain and LangGraph. . 221
Agent Skills for Offensive Security.. . . . . . . . . 240
Summary. . . . . . . . . . . . . . 251
Chapter Review Questions.. . . . . . . . . . . 252
Answers to Review Questions. . . . . . . . . . 254
Chapter 9 AI Agents for Business Leaders and Project Managers.. . . . 255
Understanding AI Agents from a Cybersecurity Leadership Perspective.. . 256
Strategic Applications of AI Agents for Cybersecurity Leaders.. . . . 256
AI Agents for Cybersecurity Project Management.. . . . . . 260
Case Studies.. . . . . . . . . . . . . 262
Summary. . . . . . . . . . . . . . 265
Chapter Review Questions.. . . . . . . . . . . 266
Answers to Review Questions. . . . . . . . . . 267
Chapter 10 Securing Agentic AI Systems. . . . . . . . . . 269
The Unique Security Challenges of AI Systems. . . . . . . 269
The Coalition for Secure AI (CoSAI).. . . . . . . . . 273
The CoSAI Risk Map: Understanding AI Security Risks.. . . . . 273
Security Controls for AI Systems.. . . . . . . . . . 281
AI Supply Chain Security. . . . . . . . . . . 289
AI Incident Response Framework.. . . . . . . . . 293
Agentic AI Architecture Security Patterns. . . . . . . . 298
Provider Versus Consumer Responsibilities. . . . . . . . 301
Regulatory and Compliance Considerations. . . . . . . 303
Case Studies and Lessons Learned.. . . . . . . . . 304
MITRE ATLAS.. . . . . . . . . . . . . 306
OWASP Top 10 for Agentic Applications. . . . . . . . 308
The MAESTRO Threat Modeling Framework.. . . . . . . 312
Summary. . . . . . . . . . . . . . 326
Chapter Review Questions.. . . . . . . . . . . 327
Answers to Review Questions. . . . . . . . . . 329
End-of-Chapter Exercises.. . . . . . . . . . . 332
9780135589854, TOC, 4/22/26
About the Author :
Omar Santos, a cybersecurity leader and Distinguished Engineer at Cisco, is renowned for his expertise in ethical hacking, vulnerability research, AI security, and incident response. He has led the DEF CON’s Red Team Village, co-chaired the Coalition of Secure AI (CoSAI), and is a prolific author. Omar's contributions have significantly advanced technology and global security standards.
