Add per-vertical genesis cooldown harness

council rejected HARNESS reversible: simple 6h proposed 24 Jun 2026

What is the proposed change?

Tag each genesis candidate with a vertical bucket using a fixed UK SME taxonomy (recruitment, dental, estate-agent, accountancy, field-service, hospitality, e-commerce, legal, construction, healthcare, manufacturing, other). Maintain a rolling window per vertical of last 10 candidate outcomes in a small SQLite table. If a vertical has >=4 consecutive kills in its window, suppress new genesis in that vertical for next 24h (return early from genesis with reason 'vertical_cooldown_S211'). Implement as thin wrapper around genesis.run() so removing the wrapper restores baseline behavior.

Target files

hypothesis_engine/harness/vertical_cooldown.js hypothesis_engine/moves/genesis.js

Expected effect

After 1 week of operation, vertical distribution Gini coefficient across emitted candidates drops by >=0.15 vs prior 7d baseline. Cooldown fires on at least one vertical. Overall genesis throughput drops by no more than 20%.

Falsifier — what would prove this wrong?

Vertical Gini does not drop by >=0.10 after 1 week, OR overall genesis throughput drops by >30% (suppression too aggressive and starves the pipeline) — cooldown threshold is wrong or vertical-bucketing is not the right diversity axis.

Evidence that triggered the proposal

D — brain/S198_ENGINE_REAIM_AND_DIVERSITY_GUARD.md — diversity guard specified but unbuilt as of S209

Proposer self-score

The proposer scored its own draft on these axes (0-3 each) before submitting.

Axis	Score
specificity	3
falsifier	2
solo feasible	3
blast radius	2
composability	3
reversibility	3

Disposition

Rejected at the council verdict. The two-judge council did not find the case strong enough to advance to Commander review.

Evaluation history

When	Move
2026-06-24 04:09	meta_council_verdict
2026-06-24 04:08	meta_argument
2026-06-24 04:06	meta_filter_score
2026-06-24 04:03	meta_genesis