Add v2_a11 urgency_event_named shadow axis to filter_score.js

Add migration s160_add_v2_a11_urgency.js that ALTERs the hypotheses table to add v2_a11_urgency_event_named INT column (default NULL). In filter_score.js highSystem and lowSystem adversarial prompt strings, insert a rubric block for v2_a11 alongside the existing 5 v1 axes. Rubric: Score 3 — hypothesis names a specific external forcing function with a date-bounded trigger: a named compliance regulation + enforcement date, a named contract renewal cadence + typical window, or a named financial consequence (OPEX waste, revenue leakage) the buyer is actively incurring. The forcing function must be external to the vendor relationship and must not be an adjective ('urgent' does not qualify; 'GDPR audit deadline Q4 2026' does). Score 2 — urgency is implied by workflow cadence or category structure (pre-send legal review implies deadline pressure; SLA breach implies consequence) but no named event is cited. Score 1 — pain is real and ongoing but purely analytical or optimization-driven; buyer can defer indefinitely without a named consequence. Score 0 — hypothesis is a dashboard, taxonomy normalizer, or analytics product where adoption has no external forcing function. Anti-gaming instruction appended to both highSystem and lowSystem: 'Score 3 requires a named external event, not a claim of urgency.' Because v2 is not yet wired into the composite graduation formula, v2_a11 scores are captured and stored for analysis without affecting current graduation thresholds. No threshold recalibration required at this stage.

RevOps Objection Taxonomy Normalizer shape (taxonomy/analytics, CRM-integrated, no named urgency event) scores 0-1. hyp-2026-05-06-ec4507 (Support Escalation with SLA breach consequences and renewal triggers) scores 2-3. Retrospective application to 43 S157-scored candidates shows bimodal clustering: ROBUST 5/5 candidates cluster at 2-3, FRAGILE and unanimously-killed candidates cluster at 0-1.

Apply axis scoring retroactively to the 43 S157 candidates using their stored descriptions. Required outcomes: (a) score distribution is bimodal — at least 40% of candidates at 0-1 and at least 30% at 2-3; (b) none of the 5 ROBUST 5/5 candidates scores 0-1; (c) at least 3 of the 4 FRAGILE candidates score 0-1. If distribution is uniform (fewer than 20% at 0-1 or fewer than 20% at 2-3), the rubric is not discriminating and must be redesigned.

Rejected by filter_score. The proposal did not meet the bar for specificity, falsifiability, or solo-feasibility.