docs: free proof surface recorded (/record, hollow-preserving, CLV honest-absent)

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QJs13VsyiSKYQP6rj3NNmc

specs/STATE.md

@@ -1072,6 +1072,39 @@
 > silent-false harness bug, and the same lesson: a verification path can lie in either direction.
 > **FLOOR: 319 suites / 3,970 tests green, web build exit 0.**
 
+> ## 📊 FREE PROOF SURFACE SHIPPED 2026-07-31 — `/record`, tier-record-forward, honest CLV panel
+> **Presentation over existing endpoints** — `src/` untouched (diff empty): no grade, model or ledger
+> change. Pricing/migration = Builds 2/3.
+> **TIER-RECORD-FORWARD.** `/record` reads the canonical public aggregates (`/api/accuracy` +
+> `/api/ledger/model`) and prints them **as-is**: **B 60% n512 · C 57% n413 — C ships honestly BELOW
+> B** · **A (n2), D (n5), F (n5) render HOLLOW with their real sample instead of a rate**. Sport
+> slicing is **client-side** because the endpoints ignore `?sport=` (verified: mlb 526 @62%, wnba 411
+> @54% from the `sports` map).
+> **🔴 THE LOAD-BEARING RULE, test-locked:** where the source withholds a percentage it **stays null**.
+> **A is 1/2, so 50% is derivable — a test asserts we do NOT derive it**, because the API withheld it
+> on purpose (n < 20).
+> **CLV IS AN HONEST ABSENCE, NOT A NUMBER.** `beat_close_pct` is null because
+> `clvCaptureReliable()` hasn't passed. The panel reads **"NOT PUBLISHED YET"** and explains that any
+> percentage printed today **"would be measuring our collection gaps as much as our edge"**; it
+> surfaces the accruing sample (937) but **no rate**. **Tests assert the panel never falls back to
+> `clv_beat/clv_sample` (34/937 = 3.6%) and that the serialized panel contains no "3.6"** — that
+> number is computable and would be WRONG, which is the exact fabrication this surface exists to
+> refuse. Built to receive a real number later without a redesign.
+> **HELD AND NAMED ON THE PAGE rather than faked:** calibration + accuracy-over-time are absent
+> because there is **no honest source** (no claimed-vs-actual endpoint; `window_days` fixed at 30 with
+> no series) — and the page says so, **and says it is "not because they are unflattering."**
+> **A page-level test asserts NO hard-coded percentage exists in the markup** (so no figure can drift
+> from the aggregate) **and that the page never touches `/api/snapshot` or itemized rows** — the
+> Build-1 gate holds, exploit stays dead.
+> **✅ FINGERPRINT — CACHE-BUSTED** (deploy boundary captured 404→200): page served ✓ · CLV
+> honest-absent ✓ · names the guard ✓ · held pieces named ✓ · **no "3.6" anywhere ✓** · **no
+> hard-coded pct in the HTML ✓**. Canonical data the page renders, verified cache-busted: **A HOLLOW
+> n2 · B 60% n512 · C 57% n413 · D HOLLOW n5 · F HOLLOW n5**; sports mlb 62 / wnba 54.
+> **⚠️ ONE ITEM NOT SERVER-VERIFIABLE:** the hollow-explanation line is **client-conditional**
+> (`loaded && rows.some(hollow)`), so it is absent from the SSR HTML by design and could not be
+> confirmed by curl → **Chrome audit** (along with the rendered bars and the sport tabs).
+> **FLOOR: 320 suites / 3,986 tests green (16 new), web build exit 0.**
+
 - **Redirect EXISTS + WIRED:** `closingCapture.buildCaptureRows`→`closing_captures` (append-only,
   provenance: captured_at/book/line_type/both-prices/missed_reason) via `intradayRefreshService:221`
   + internal endpoint; `ledgerService.attachClosingProb`→`closing_prob` (de-vigs both raw sides,