← back to lesson
Coding JARVIS · stakeholder reference

2 · Adversarial Testing

The Guardrails

Aim. Learn to probe a system for failure, attacking it with adversarial prompts, to produce a hardened agent.

Format shape. red-team-an-agent · drill

4 stages · 9 scenes

1arcade

  1. Connections
  2. hook Red Team Race
  3. hook The Literal Genie~18m

2dossier

3workbench

4larp

  1. Cold Open
  2. Brief
  3. Rubric
  4. Roles
  5. Execution
  6. Boss Fight