Re-arm the UK QRF tail-concentration gate on independent draws #1277

The UK QRF tail-concentration gate (uk_qrf_tail_concentration) counts rows. Its thresholds were armed on design-weight measurements of an earlier spine, but the frames it now runs on duplicate records:

• the CGT clones and the SPI support copies on every UK frame;
• the K geographic clones on the local line.

A top-100 share is therefore measured over copies, not over independent draws. Thin instruments can cross the carrier floor on copies alone, and a cloned pool dilutes a genuine concentration.

Three SPI instruments now sit behind national-line reviewed exclusions in qrf_tail_reviewed_exclusions.json (PR #1242):

They are rare on the SPI tape and look genuinely concentrated whether counted by row or by independent record, so they are not the #462 donor-tail broadcast.

To do before 2027-01-14:

1. Measure the gate on independent draws: collapse CGT clones, SPI copies and geographic clones to their source record.
2. Re-arm top_k, max_top_share and min_nonzero_records on that measurement.
3. For each of the three instruments, decide whether it stays behind a reviewed exclusion as a concentrated instrument or clears under the re-armed gate.
4. Renew or retire the three entries accordingly.

Licensed figures stay out of this issue; measurements go to the owner privately.

Owner: @juaristi22. Raised in review of #1242.