AI Agent Security handbook: Strategies for Protecting Autonomous AI Agents from Prompt Injection, Memory Attacks, and Emergent Behaviors - Couverture souple

Livre 1 sur 2: Agentic AI Security Mastery

Hartmann, Leo

 
9798277568651: AI Agent Security handbook: Strategies for Protecting Autonomous AI Agents from Prompt Injection, Memory Attacks, and Emergent Behaviors

Synopsis

Autonomous AI agents can execute complex workflows, maintain persistent memory, and interact intelligently with their environment—but with great capability comes significant risk. The Agent Security Playbook provides a hands-on, practical guide to securing LLM agents against the most common and emerging threats.
Inside, you’ll learn how to:

  • Develop agent-specific threat models and identify attack surfaces
  • Engineer safe, schema-bound prompts and orchestrate tool access
  • Implement memory integrity checks, anomaly detection, and write controls
  • Embed real-time safety critics and policy enforcement into agent reasoning
  • Conduct adversarial testing and red teaming for proactive defense
This volume equips AI professionals with foundational and tactical strategies to build autonomous agents that are both powerful and secure, serving as the cornerstone for the advanced techniques explored in Volume 2.

Les informations fournies dans la section « Synopsis » peuvent faire référence à une autre édition de ce titre.