A flight simulator for the human side of AI agents: practice Allow, Ask, or Block before the consequences are real.
Simulates agent actions and scores the user's Allow, Ask, or Block choice against the consequences.