Track 15: Frontier and Advanced, Agent Security and Red Teaming

How Agents Get Attacked: The Risky Combination

Spot the risky combination in any agent design.

Status: Coming soon BeginnerVideo lessonAbout 12 min
Get notified on YouTube Udemy certificate: coming soon

Course

AI Security: Prompt Injection, OWASP LLM Top 10 & Guardrails

Udemy: coming soon

Watch first

Then try

In this module

  1. How Agents Get Attacked: The Risky Combination
  2. Agent Defences That Hold Against Prompt Injection
  3. MCP Security Lab: Tool Poisoning, Tokens, OAuth Scopes
  4. Red-Team Agents With PyRIT, garak and promptfoo

The course this lesson belongs to