Securing the Autonomous Supply Chain: Defending the new attack surface of agentic operations across the physical-digital convergence
暫譯: 保護自主供應鏈：防禦實體與數位融合下代理式營運的新攻擊面

The autonomous supply chain is the emerging attack surface of the next decade. When AI decisions can drive physical action, the execution perimeter becomes the new security boundary. Explore the threat chain across the physical-digital convergence through a practical approach to implementation, with 17 customizable artifacts from the ASCR blueprint to turn security strategy into action.

Key Features:

- Map the new attack surface from agentic decisions to physical consequences

- Secure the agentic execution chain from decision to physical action

- Turn security strategy into action with practical, customizable artifacts

Book Description:

As agentic systems progress from human-approved decisions to semi-autonomous and autonomous execution, the security target is changing. The critical question is no longer only how AI can be compromised, but how to secure the execution boundary when an agentic system can acquire the authority to affect physical operations. Agentic operations create a new attack surface across the Physical-Digital Convergence, where digital intelligence can drive physical action through autonomous execution. The Execution Perimeter defines where execution authority must be validated before decisions become physical actions.

This book traces the Threat Chain from AI reasoning through connectivity and actuation to physical consequences. It moves from threat analysis to securing, validating, deploying, monitoring, and responding to agentic systems, using four threat scenarios and integrated attack models to demonstrate the security challenges of autonomous operations.

By the end of the book, you will learn how to secure agentic execution chains, validate execution authority, and build resilient security architectures across the Physical-Digital Convergence. The ASCR Blueprint provides a practical path from strategy to implementation, with 17 reusable artifacts spanning risk intelligence, technical enforcement, operational response, and governance.

What You Will Learn:

- Trace the threat chain from AI reasoning to physical consequence

- Identify and intercept attacks across the agentic execution chain

- Secure MCP gateways, agent identity, lineage, telemetry, and evidence

- Use digital twins to test agents before physical execution

- Build an Agentic Security Operations Center (ASOC)

- Apply the Secure Agentic Development Lifecycle (SADLC)

- Align security with EU AI Act, NIS2, IEC 62443, and NIST AI RMF

- Apply the ASCR Blueprint and 17 artifacts across risk and response

Who this book is for:

This book is for security architects, CISOs, security and technology leaders, AI/ML and application security engineers, OT/ICS engineers, cloud security professionals, and compliance practitioners responsible for securing, governing, or auditing agentic systems in production. Consultants will also benefit from the reusable frameworks and customizable artifacts throughout the book.

Basic security architecture knowledge is recommended; familiarity with AI/ML and OT/ICS is helpful. Cross-domain context is established where required.

Table of Contents

- The Agentic Revolution: From Copilots to Autonomous Execution

- The Physical AI Security Landscape - The Cyber-Physical Systems Threat Chain

- The Agentic Threat Matrix - Attacks on Intelligence and Coordination

- The Hardware and Edge Attack Surface - Attacks

- Securing Models, Data, and Decision Provenance

- The Secure Agentic Development Lifecycle

- Identity Continuity and Access Control Across the CPS Stack

- The Agentic Security Operations Center (ASOC)

(N.B. Please use the Read Sample option to see further chapters