Add pre-filter gate detecting v2_backfill_orphan_S148 condition

filter rejected GATE reversible: simple 6h proposed 17 Jun 2026

What is the proposed change?

Create new gate module gates/backfill_orphan_check.js. Gate runs between genesis and filter_score. It inspects the candidate's evidence claims and flags any claim whose 'source' field is null, 'self-asserted', or references a corpus item that was added AFTER the genesis prompt was constructed (orphan-backfill pattern). Gate emits FAIL on any orphan claim and short-circuits the candidate to kill_reason='v2_backfill_orphan_S148_prefilter' WITHOUT spending argument/judge tokens. Wire from cycle.js: call gate.check(candidate) immediately after genesis returns.

Target files

hypothesis_engine/moves/genesis.js hypothesis_engine/gates/backfill_orphan_check.js

Expected effect

Of the 7 v2_backfill_orphan_S148 kills in Corpus E, all 7 would be caught at the pre-filter stage. Argument/judge spend on these candidates drops to $0. Total saved per 100-run batch: ~$0.40 (7 candidates * argument+judge cost).

Falsifier — what would prove this wrong?

Backtest gate on the 7 logged orphan kills. If fewer than 6 of 7 trip the gate, the detection logic doesn't match the actual orphan signature and the gate is misdesigned. Also: if gate fires on >5% of currently-graduating KEEP candidates, false positive rate is too high.

Evidence that triggered the proposal

E — kill_reason_distribution: v2_backfill_orphan_S148 = 7 occurrences, top recurring failure
D — brain/META_ENGINE_PHASE_1_SPEC.md — pre-filter gate pattern already used elsewhere

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	2
reversibility	3

Disposition

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

Evaluation history

When	Move
2026-06-17 04:04	meta_filter_score
2026-06-17 04:03	meta_genesis