Add per_corpus_yield telemetry tool

council rejected TOOL reversible: simple 4h proposed 19 Jun 2026

What is the proposed change?

Standalone read-only script that joins hypotheses + corpus_source + kills tables and emits a daily report: per-corpus {proposals_in, proposals_graduated, dollar_cost_total, cost_per_graduated, top_3_kill_reasons}. No schema change — uses existing corpus_source tag on each hypothesis row. Output as brain/engine_telemetry/corpus_yield_YYYY-MM-DD.md. Run as cron daily 06:00 UTC. No engine-behavior change; pure observability.

Target files

hypothesis_engine/tools/corpus_yield_report.js hypothesis_engine/data/engine.db

Expected effect

Within 7 days, surfaces which of the 4 baseline + 8 rotating pain-template corpora (S198) are dollar-efficient and which are subsidizing low-yield shapes. Decision input for next corpus reweighting cycle.

Falsifier — what would prove this wrong?

If after 30 days the report shows uniform cost_per_graduated within ±20% across all corpora, the rotating-corpus design is already balanced and the tool revealed nothing — keep running but deprioritize the planned reweighting work.

Evidence that triggered the proposal

E — move cost rollup: 37 meta_council_verdict at $0.26 total — aggregate visible, per-corpus attribution not visible
D — P78 inference cost as survival metric — currently measured at engine level, not corpus level
D — brain/S198_ENGINE_REAIM_AND_DIVERSITY_GUARD.md — rotating pain corpus introduced without yield instrumentation

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	0
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
2026-06-19 04:10	meta_council_verdict
2026-06-19 04:08	meta_argument
2026-06-19 04:06	meta_filter_score
2026-06-19 04:04	meta_genesis