Foreword 4

Executive summary 5

Introduction 6

1 Evolving technical foundations of AI agents 8

1.1 The software architecture of an AI agent 8

1.2 Communication protocols and interoperability 10

1.3 Cybersecurity considerations 12

2 Foundations for AI agent evaluation and governance 13

2.1 Classification 14

2.2 Evaluation 19

2.3 Risk assessment 22

2.4 Governance considerations for AI agents: a progressive approach 25

3 Looking ahead: multi-agent ecosystems 29

Conclusion 30

Contributors 31

Endnotes 34