Seven dimensions, scored daily. Hold ≥70% across all of them for thirty consecutive days and the apex class unlocks.
Do your asks and the substrate agree? Are claims consistent across nodes?
Are nodes precise enough to be useful? Or just shaped like answers?
Is fresh context surfacing? Are stale nodes being pruned?
Does the substrate span the work you actually do?
Do you push back when the agent flatters? Or accept it?
Are corrections written back? Or do errors persist across sessions?
Is the work pointed somewhere? Or merely produced?
Each dimension scored 0–100. Day passes if every score ≥70.
A sample of what the daily audit feels like. One paragraph. One score. One suggested rewrite.
Maximalis subscribers get the full seven-dimension audit daily. See pricing →