What this mock holds

Symmetric difference against wave0/asset_inventory.csv (a set, never a count)

In the inventory, not on the map — 0:

• — (empty)

On the map, not in the inventory — 10 (the ten region anchors: compositions, not assets; printed so they can never pass as inventory):

• canvas:C1
• canvas:C2
• canvas:C3
• canvas:C4
• canvas:C5
• canvas:C6
• canvas:C7
• canvas:C8
• canvas:UN
• canvas:WS

Every node href against :8798

Resolve 200: 238 of 279 distinct hrefs · gated 403: 40 · missing: 1. Not 200:

• 40 gated (403 — the internal gate; WORKSHOP's 'gated, never hidden'; kept as the real route): /bankruptcy/ranked-surface, /drafts, /drafts/courts, /fusions, /fusions/courts-combo-venue-atlas, /lab/design-system, /lab/distress-ip-watch, /lab/ingest-accrual, /lab/kit, /lab/title-report, /lab/vintage-custody, /lab/workspace, /practice, /preview/console, /studio, /studio/adjudicate, /studio/authority, /studio/book, /studio/brands, /studio/censoring-atlas, /studio/claim-feasibility, /studio/climate, /studio/counsel, /studio/deliverables, /studio/drug-cliff, /studio/entity, /studio/entity/section/constellation/offshore, /studio/feed-horizons, /studio/forum, /studio/graph, /studio/graph/frag/table, /studio/hub/ledger, /studio/judge, /studio/mdl, /studio/morgue-quarterly, /studio/negative-certificate, /studio/regime-breaks, /studio/scotus, /studio/termination-windows, /studio/value
• 404 /v3/workspace?view=capital — the inventory's own lives_today route; kept and printed rather than re-pointed

Resolved 200 through a redirect (the final route is what serves the bytes):

• /mdl
• /sentencing/ussc
• /verticals/anda
• /verticals/bankruptcy
• /verticals/ptab

Figures — where they come from

Every lit figure is read off a page the V3 app printed at build time (.v3r-figure__label · .v3r-figure__num · .v3-gate on the 46 rooms; the door hero measures). Card 01's survivorship gap is not printed today: /almanac answered 200; states ['REFUSED', 'refusal']; refusal cells 2; card01 binding DARK on the carrier: True — so TIME's figure is the appellate pendency median the matter door prints, and the gap is a hatched node with the carrier's sentence. Nothing on this page is computed from data; nothing is predicted.

Scripted pan + zoom — frame time

Not yet run. Run the scripted pan + zoom (also runs with ?bench=1).

The 25 checks (§8) — self-score on this page

What it does not prove

• §7.4's WebGL light field and per-frame grain — the ground here is the two-gradient fallback with static grain (allowed; not the shader).
• deck.gl edges — deliberately not loaded (1.6 MB, no served route uses it, a canvas node is invisible to the crawl guard); edges are one SVG inside the transformed layer. Stated choice, not a proof that WebGL edges would be worse.
• Card 01 / 07 / 13 figures on the map — their carrier prints withheld columns today; the map cannot light what the app does not print.
• The 3D courthouse scene inside PLACE's canvas panel (M2's object) — the panel here previews composition with real room figures only.
• Frame-time p95 under hardware Metal — measured here under whatever GPU the reviewer's browser has; the archived number is SwiftShader/headless.
• The region assignment is a ruling (§3.3/§3.6) encoded in build_estate_nodes.py; the node SET is the census. A reviewer who moves a room moves it in that dict, never on the page.

Counts at build: {"nodes": 439, "by_tier": {"1": 10, "2": 66, "3": 363}, "by_state": {"lit": 58, "hatched": 288, "dim": 92, "empty": 1}, "by_region": {"C1": 54, "C2": 50, "C3": 18, "C5": 27, "C4": 33, "C6": 25, "C7": 26, "C8": 103, "WS": 92, "UN": 1}, "edges": 737, "edges_by_kind": {"composes": 429, "same-entity": 207, "receipt": 44, "what-else": 57}, "inventory_rows": 429, "index_rows": 836}