Veto & deterioration list — most flags first (veto, then deterioration). Credibility is a promise-keeping + deterioration-veto record — falsified as a return factor, not a ranked long list. Record uses measurable items only (D61): guidance accuracy, quantification %, the ⛔ veto, and deterministic deterioration. The last three ·AI columns are a model read shown for context — NOT ranked. A name is unproven until its promises resolve. Pilot — the historical backfill accrues ≈18 concalls/day via the nightly cron; open any name for its full dossier. Corpus provenance (guardrail-#8 disclosure): ~98.6% of transcripts were discovered via Screener.in links (legacy path, frozen; BSE-primary migration in progress) — extraction and settlement run on the primary documents themselves.