Add forecaster-grade pre-graduation gate at filter→argument boundary

filter rejected GATE reversible: simple 5h proposed 11 Jun 2026

What is the proposed change?

Between filter_score pass and argument move, call forecaster/grade.js with the candidate as input. If forecaster confidence (calibrated S184 score) <0.3 OR source-separated evidence count <2, kill with reason='forecaster_low_confidence'. The forecaster already exists; this gate is a single import + 6-line check. No new model calls beyond what forecaster already does.

Target files

hypothesis_engine/cycle.js forecaster/grade.js

Expected effect

Candidates that pass filter but die at argument due to thin evidence drop by ≥30% (estimated from last 14 days of engine.db). Argument move cost-per-graduated-candidate drops proportionally.

Falsifier — what would prove this wrong?

Run gate in shadow mode (log decision but don't kill) for 100 cycles. If forecaster-would-kill candidates show no different argument-survival rate than forecaster-would-pass candidates (Δ <10pp), forecaster is not load-bearing at this boundary and gate is noise.

Evidence that triggered the proposal

D — brain/S184_FORECASTER_V1_CALIBRATION.md
E — argument-move kill-reason distribution: 'thin-evidence' subset

Proposer self-score

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

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

Disposition

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

Evaluation history

When	Move
2026-06-12 04:42	meta_filter_score
2026-06-11 04:03	meta_genesis