Truth-label register — Executive Demo r1.0
r1.1 corrections (2026-09-14). Copied from
mizoki-exec-demo-r1.0.zip(docs/, zip sha25698fbd2fc…1282) and then trued againstmainon the day it landed, as the register itself demands: the page moved to/media/executive-demo(owner placement ruling 2026-09-13; slug renamed to/media/demo2026-09-14 by owner call, the old slug 308s there;/demo/executiveanswers 404, test-pinned); the cells chip carries §B.6's observe-only qualifier and the close table's ORACLE row reads PARTIAL; the F4 row reads LIVE with its earned-authority qualifier (owner-delegated ruling 2026-09-13,docs/CANON_STATUS.md); the close table is checked against the MACHINE copy of §B.6 (contracts/mizoki_contracts/canon_status.py) bytests/test_demo_executive.py. Rows marked (r1.1) changed; every other row is the r1.0 text.
Every status chip and figure on /media/demo (r1.1: /media/executive-demo; r1.0: /demo/executive), with the canon source it was taken from. If main disagrees with a row on the day you land this, fix the page, never the source.
| Surface on page | Label shown | Canon source |
|---|---|---|
| Profit-Leak Report, waterfall, dosage tile | Net yield · IN BUILD · writeback OFF | Whitepaper r2.0 §6, §7; OFFERING_MAP §B.6 "Net Yield lane — IN BUILD" |
| Session replay header | Cells 33–36 · LIVE · observe-only (r1.1) | Whitepaper r2.0 §3; OFFERING_MAP §B.1 fleet paragraph; §B.6 "ORACLE intent cells 33–36 — PARTIAL, observe-only / shadow" |
| Forecast gauge, intent facts | Intent Engine v2 / I-02 · IN BUILD · shadow | Whitepaper r2.0 §4; OFFERING_MAP §B.6 |
| O-1 moment, bright-lines panel | Privacy lock · LIVE · owner ruling | Whitepaper r2.0 §3.2; ruling O-1 |
| Holdout registration, lift chart | Causal core · PARTIAL; X-/DR-learner; DoWhy | Whitepaper r2.0 §5; OFFERING_MAP §B.6 "Causal proof core — PARTIAL" |
| Ghost bid method button (disabled) | PROPOSED | Whitepaper r2.0 §5.2 |
| Causal Credit Ledger | unified.causal_credit_ledger · PARTIAL | Whitepaper r2.0 §5.3 |
| Dosage curve | continuous dosage · PARTIAL · advisory | FEATURE_CATALOG r2.1 Part 2A row 9; DEMO trace act 4 (r_learner_local_marginal_moment_wls_v1, d* 4047.62) |
| Proposed-action panel | DCP · threshold gate · LIVE | Whitepaper r2.0 §14.3; OFFERING_MAP §B.6 row 1 |
| DEL chart dashed curve + note | clipped-ReLU shape · PROPOSED CANON, "not claimed" | Whitepaper r2.0 §8.1; Feature Coverage Matrix gap G-18 |
| DEL formula in note | 100×(0.5·pass rate + 0.3·evidence + 0.2·verification) | services/service-policy-engine/main.py:176-178 per whitepaper r2.0 §14.3 |
| Floor slider | DEL ≥ 80.0 platform floor, never lowered | Whitepaper r2.0 §8.1; OFFERING_MAP §B.6 vocabulary |
| Guardrail clamps in passport | budget ±20%, bid ±30%, 15-conversion floor | mizoki_runtime/demo_signal.py GuardrailSet; OFFERING_MAP §B.3 step 4 |
| F4 reservation gates | F4 · LIVE · pilot armed 2026-08-24 · every reservation L2-approval-gated · bounded autonomy after 2 clean cycles (0 complete); three HALTs (r1.1) | F4_GO_LIVE_2026-08-24; owner-delegated ruling 2026-09-13 (docs/CANON_STATUS.md F4 row, AGENTS.md 9.1); DEMO trace act 5 f4_gates |
| Restraint Ledger rows | illustrative composites from the Signal Factory scenarios + trace | demo_signal.py planned_actions; DEMO trace act 5 |
| Rail footer | MEASUREMENT_WRITEBACK OFF · NET_YIELD_WRITEBACK OFF · recommend-only | Whitepaper r2.0 §7; §14.4 |
| Close: autonomy gates | Brier ≤ 0.20 · AUC ≥ 0.72 · ≥ 2 cycles | Whitepaper r2.0 §10.1 |
| Close: "AUC gate cannot be scored — zero admissible forward labels → observe" | verified result for the label-loop count (605,407 scored rows, 0 admissible forward labels, measured 2026-08-10); the 0.72 line is a design target; no AUC figure on the slide (r1.1 corrected 2026-09-14: the former "AUC 0.6884, measured" was the fallback scorer on one seed and "measured" is not a TRUTH.md label) | docs/reports/LII_BACKTEST_MODEL_QUALITY_2026-08-16.md (Measured table; recommendation 1); docs/BUILD_DEBT.md INT-EVAL-1 |
| Close: status table | LIVE / PARTIAL / IN BUILD / PROPOSED rows — F4 LIVE (qualified); ORACLE cells 33–36 PARTIAL; Intent Engine v2 · Net Yield · F1 F2 F3 F5 IN BUILD (r1.1) | OFFERING_MAP v2.3 §B.6 deployment-state table; machine copy contracts/mizoki_contracts/canon_status.py (test-pinned row by row) |
| Act 1 figures | 847 / 23 / −$14,280 / 42%; SKU-DEMO-7742 4.2× · 58% · −$620 | DEMO_FIVE_ACT_TRACE_2026-08-25.json act 1 (ILLUSTRATIVE) |
| Scale selector | dollar figures × 1 / 4.8 / 24 | demo-only convenience; counts (SKUs, conversions) never scale |
Vocabulary rules honoured: "anticipatory intent", never "mind-reading" · "nervous system", never "brain" · no revolutionary/AI-powered/best-in-class · claims classed, never mixed.