What does day-two operation look like?

Deployments are usually planned through go-live and not beyond, which is why so many degrade quietly. The ongoing work is small but real: rules go stale, models shift underneath you, and consumption patterns change as adoption grows.

Review cadence is a field on each skill, so overdue rules surface as a query against the Skills Registry rather than as a memory exercise. The Interaction Log supports both drift detection — re-running known questions and comparing full assembled context — and spend attribution per skill and requester. Updates ship every two weeks, so provider and model changes arrive as scheduled adaptations rather than as surprises.