Build nested typed evidence rows directly while retaining every original field and validation. Complete serialized bytes match the prior writer on recorded d8 evidence and regression fixtures.
Fresh-head combined qualification: 459 tests passed. Cloud performance measurements do not claim device loop-rate recovery or a successful learning trial.
Keep bounded automatic learning collection through valid live calibration refinements. Retire generations on resets or invalid calibration, preserve live physical transforms and existing controller tunes, and validate fitted coefficient bounds before publication. Exercise the current controller and historical wire paths.