Daily improvement loop: scan scorecards -> propose challenger -> shadow trial -> promote.
Feedback service offline
Self-learning data unavailable.
Terminal: make up
or System Health → Start Service
Tier 1 (autonomous)
0
Prompts, exemplars, modifiers
Tier 2 (approval)
0
Routing changes
Tier 3 (report)
0
Thresholds, weights
Last Cycle
Not run yet
Daily at 02:00
Each reflection is an improvement proposal from one of the 9 learning workers.
Run the scheduler or queue learning_cycle manually to start the loop.