Leanpub Header

Skip to main content

Breaking AI

The Practical Guide to AI Red Teaming, LLM Exploitation, and Agentic Security

Breaking AI

A practical guide to testing and defending AI applications, from prompt injection and RAG to MCP tools, autonomous agents, and incident response.

Minimum price

$19.00

$29.00

You pay

Author earns

$

Also available for 1 book credit with a Reader Membership

PDF
About

About

About the Book

An AI assistant can read a document, retrieve a customer record, and propose an action. The security question is not only what the model says. It is what the surrounding system permits—and what evidence shows actually happened.

Breaking AI connects AI red teaming with enterprise security architecture. Across 31 chapters, Igor van der Burgh examines prompt injection, jailbreak evaluation, context poisoning, sensitive information exposure, retrieval security, MCP tools, agent delegation, persistent memory, multimodal inputs, and the AI supply chain. Later chapters turn those risks into assessment methods, defensible measurements, secure designs, incident response, and governance decisions.

Readers follow Northbridge Services, a fictional organization introducing a knowledge assistant and progressively more capable agents. The recurring method is practical: identify the trust boundary, state the forbidden outcome, run an authorized test, inspect the effect, improve the control, and retest legitimate use.

The companion package includes 21 executable synthetic control scenarios, regression tests, a separate optional local-model adapter, editable diagrams, and reusable field worksheets. The offline labs test explicit application controls; they do not pretend to prove that a language model is universally safe.

Written for architects, security engineers, AI developers, incident responders, and technical leaders, Breaking AI offers a balanced route from understanding an attack to engineering a defense that can be examined and challenged.

Share this book

Author

About the Author

Igor van der Burgh

Igor van der Burgh is a Lead Solution Architect at Citrix with almost 25 years of experience in

information technology. His career has centered on end user computing, progressing from consultant to

field architect and then to his current role.

In recent years, Igor has increasingly focused on cybersecurity and artificial intelligence. That interest

has led him to explore how AI systems can be manipulated, how their data and connections create

exposure, and how technical choices affect organizations and the people who depend on them.

This book grew out of his own learning journey. As Igor deepened his understanding of AI security, he

wanted to turn what he learned into an accessible resource for others. His aim is to connect explanations

with practical examples and questions that readers can apply in their own work.

Drawing on his background in consulting and architecture, Igor writes for people who need to translate

technical subjects into organizational decisions. He hopes readers will use the book to examine

assumptions, start useful conversations, and identify practical improvements, whether advising

customers or strengthening security within their own organizations.

For Igor, writing is also a way to keep learning. This book reflects his belief that knowledge becomes

more valuable when it is shared and put into practice.

Connect with Igor

linkedin.com/in/igor-van-der-burgh-5284947

Contents

Table of Contents

The Practical Guide to AI Red Teaming, LLM Exploitation, and Agentic Security Igor van der Burgh Copyright and responsible use About the Author Preface: breaking assumptions, not trust How to use this book Introduction: the Northbridge system Part I — Understanding the AI Attack Surface 1. The New Battlefield 2. Anatomy of Modern AI Systems 3. AI Threat Modeling 4. Building a Safe AI Red Team Laboratory Part II — Attacking Language Models 5. Prompt Injection: When Data Pretends to Be Authority 6. Jailbreak Testing and Safety Boundaries 7. Context Poisoning and Conversation Manipulation 8. Sensitive Data Exposure and Information Boundaries 9. Model Extraction, Inversion, and Privacy Part III — Exploiting the AI Ecosystem 10. RAG and Vector Database Security 11. MCP, Tool Calling, and Plugin Security 12. Agentic AI: Autonomy, Delegation, and Privilege 13. Multi-Agent Systems and Cross-Agent Trust 14. AI Supply Chain, Model Artifacts, and Infrastructure Part IV — Advanced and Emerging Threats 15. Multimodal Attacks: Images, Audio, Video, and Documents 16. Persistent Memory Poisoning 17. AI-Assisted Social Engineering and Verification 18. Autonomous Attack Chains and Cascading Failures Part V — Practical AI Red Teaming 19. Red Team Methodology: From Permission to Evidence 20. Testing Tools and Frameworks 21. Designing Adversarial Evaluations 22. Measuring Exploitability, Impact, and Control Effectiveness 23. Reporting Findings and Building an AI Security Scorecard Part VI — Engineering AI Defenses 24. Designing Secure AI Architectures 25. Identity, Authorization, Sandboxing, and Tool Permissions 26. Guardrails, Validation, and Their Limits 27. AI Security Monitoring, Detection, and Incident Response 28. Continuous AI Security Operations Part VII — Governance and the Future 29. Frameworks, Standards, and Evidence 30. Regulation, Compliance, and Supplier Assurance 31. The Future of Offensive and Defensive AI Security Appendix A. The AI Red Team Field Checklist Appendix B. Laboratory Workbook Appendix C. Reusable Worksheets and a Worked Finding Appendix D. Chapter and Framework Crosswalk Appendix E. AI Incident Response Cards Appendix F. Glossary Appendix G. Exercise Review and Discussion Guide Appendix H. Implementation Patterns and Production Extensions Reading Index References

Get the free sample chapters

Click the buttons to get the free sample in PDF or EPUB, or read the sample online here

The Leanpub 60 Day 100% Happiness Guarantee

Within 60 days of purchase you can get a 100% refund on any Leanpub purchase, in two clicks.

See full terms...

Earn $8 on a $10 Purchase, and $16 on a $20 Purchase

We pay 80% royalties on purchases of $7.99 or more, and 80% royalties minus a 50 cent flat fee on purchases between $0.99 and $7.98. You earn $8 on a $10 sale, and $16 on a $20 sale. So, if we sell 5000 non-refunded copies of your book for $20, you'll earn $80,000.

(Yes, some authors have already earned much more than that on Leanpub.)

Learn more about writing on Leanpub

Free Updates. DRM Free.

If you buy a Leanpub book, you get free updates for as long as the author updates the book! Many authors use Leanpub to publish their books in-progress, while they are writing them. All readers get free updates, regardless of when they bought the book or how much they paid (including free).

Most Leanpub books are available in PDF (for computers) and EPUB (for phones, tablets and Kindle). The formats that a book includes are shown at the top right corner of this page.

Finally, Leanpub books don't have any DRM copy-protection nonsense, so you can easily read them on any supported device.

Learn more about Leanpub's ebook formats and where to read them

Write and Publish on Leanpub

You can use Leanpub to easily write, publish and sell in-progress and completed ebooks and online courses!

Leanpub is a powerful platform for serious authors, combining a simple, elegant writing and publishing workflow with a store focused on selling in-progress ebooks.

Leanpub is a magical typewriter for authors: just write in plain text, and to publish your ebook, just click a button. (Or, if you are producing your ebook your own way, you can even upload your own PDF and/or EPUB files and then publish with one click!) It really is that easy.

Learn more about writing on Leanpub