What you'll learn

This path is the sequel to the AI Security Path, built around the OWASP Top 10 for Agentic Applications. Learners start with how agents plan, use tools, and delegate, then work through all ten risks (ASI01 to ASI10) in the order an attacker would meet them: hijacking an agent’s goal, abusing its tools and identities, exploiting its code execution, memory, and communication, and finally the failures that spread when things go wrong. Every module pairs the attacks with the defences that stop them, and finishes with hands-on challenge rooms built on the agentic platform.

5 Modules24 RoomsIntermediate

Module 1

Agentic AI Fundamentals

Early accessAgentic Fundamentals Early accessAgentic Threat Modelling Early accessAgentic Security Monitoring Early accessAgentory

Module 2

Agent Hijacking

ASI01: Agent Goal Hijack

ASI01: Indirect Injection in Agentic Systems

ASI01: Defending Against Agent Goal Hijack

Off the Brief

Proxy Meeting

Module 3

Agent Components & Authority

ASI02: Tool Misuse & Exploitation

ASI03: Agent Identity & Privilege Abuse

ASI04: Agentic Supply Chain Compromise

Securing the Tool and Identity Layer

Privilege Escalation via Agent Chain

Module 4

Agent Runtime & Memory

ASI05: Unexpected Code Execution

ASI06: Memory & Context Poisoning

ASI07: Insecure Inter-Agent Communication

Defending the Persistence Layer

The Trust Cascade

Module 5

Rogue Agents & Systemic Failure

ASI08: Cascading Agent Failures

ASI09: Human-Agent Trust Exploitation

ASI10: Rogue Agents

Agentic Incident Response

The Rogue Pipeline