Agentic AI Security and Guardrails

Authority, evidence and safe action

Rajesh Ranjan Mahapatra

Threat models, identity and delegation, prompt injection, data protection, retrieval and memory: follow the boundary between information and authority through an agent system.

Browse all books

This book is being prepared for publication on Kindle. The free online edition and downloads have been withdrawn in the meantime.