What does it mean that document your agents?

Example: You have a morning routine agent: when the alarm goes off, you make coffee, check your calendar, and plan your top three priorities. This agent has run reliably for two years — but it lives entirely in your head. One stressful week, you skip the planning step. Then you skip it again. Within a month, the agent has silently degraded and you don't notice until you're regularly arriving at work reactive instead of prepared. Had you written it down — trigger, conditions, actions — the degradation would be visible in the gap between the document and your behavior.

Try this: Pick one agent you already run — a decision rule, a recurring process, a behavioral protocol. Write it down in this format: (1) Name, (2) Trigger — what activates it, (3) Conditions — when it applies and when it doesn't, (4) Actions — the specific steps, in order, (5) Success criteria — how you know it worked. Time yourself. If it takes more than ten minutes, the agent is probably too broad (see Agent scope should be narrow). If you can't articulate the trigger or conditions, you've discovered the agent was vaguer than you thought.