What does it mean that define success metrics for each agent?

Every agent needs a clear definition of what success looks like in measurable terms. Without operational metrics, monitoring produces noise instead of signal.

Example: You built a morning routine agent — a sequence of habits designed to start your day with intention. You 'monitor' it by vaguely asking yourself whether mornings feel good. Some days feel productive, some don't, but you can't tell whether the agent is working or you just slept well. Now define three metrics: completion rate (did you execute all steps?), start latency (how many minutes between alarm and first action?), and energy rating at 10am (1-5 scale). Suddenly you can see that your agent fires reliably on weekdays but collapses on weekends, and that start latency above 15 minutes correlates with energy ratings below 3. The agent didn't change. Your ability to see it did.

Try this: Pick one cognitive agent you currently run — a decision-making heuristic, a weekly review process, a reading protocol, anything. Write down three metrics that would tell you whether it is succeeding: one measuring whether it fires at all (reliability), one measuring whether it produces the intended output (effectiveness), and one measuring cost (time, energy, or cognitive load). Be specific enough that someone else could measure them without asking you what you mean.