What is agent monitoring?

Agent monitoring is a concept in personal epistemology: An agent that fails to fire when it should leaves you exposed to undetected problems — the silence feels like safety, but it is blindness.

Example: You built an agent that's supposed to flag when you're overcommitting your calendar. Last month you said yes to six evening events, missed three family dinners, and burned out by Friday — but the agent never fired. You didn't notice the pattern because no alarm went off. The absence of a warning felt like permission. It wasn't. Your agent's false negative rate was 100%: every situation it should have caught, it missed.

This concept is part of Phase 28 (Agent Monitoring) in the How to Think curriculum, which builds the epistemic infrastructure for agent monitoring.