test(mcp): Villa Azul + 10-agent deep verification

Builds a larger, richer house than Casa del Sol via MCP save_scene
(no injection hack), then dispatches 10 parallel verifiers across
schema, geometry, dimensions, openings, HTTP, page render,
parentage, round-trip, spatial, visual.

Villa Azul — 56 nodes, validate_scene=true, 44KB on disk:
- 15x10m building envelope (vs Casa del Sol's 12x8)
- 9 interior zones (master bed/bath, bed 2/3, shared bath,
  living/dining, kitchen, entry hall, corridor)
- 10 doors + 12 windows (all cut successfully)
- 4 exterior zones (pool 8x4 + basin slab at -2m, outdoor kitchen,
  driveway, back patio)
- 5 rail-style fences (vs Casa del Sol's privacy) with 2m entrance gap

Verification: 108 checks, 104 PASS, 4 findings:
- V1 schema: 56/56
- V2 geometry: 7/7 (perimeter closes, interior T-junctions, no
  zone overlaps, fence gap verified)
- V3 dimensions: 13/13 zone areas exact (1 spec mismatch on site
  polygon default, not a build bug)
- V4 openings: 22/22 dimensional fit, surfaced a tool gap in
  cut_opening (no adjacency check) + my build packed too tightly
- V5 HTTP: 10/10 (GET/PUT/PATCH/DELETE/HEAD, If-Match conflicts)
- V6 page: 14/14 (/scene/:id 81KB, /scenes 20KB, 404 fallback)
- V7 parentage: surfaced CROSS_CUTTING §2 site->building->level
  parentId=null (pre-existing in core's loadScene)
- V8 round-trip: 10/10 byte-equal, duplicate_level -> 110 nodes
- V9 spatial: 12/12 (find_nodes, measure, constraints resource)
- V10 visual: HTML fallback (Chrome extension disconnected during
  run); API layer intact

Follow-up tracked: `cut_opening` should check opening-adjacency on
the same wall (minimum gap) to catch tight packing during patch
construction. Currently returns success and relies on the UI to
visualise the overlap.

Live at http://localhost:3002/scene/a6e7919eacbe.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Adrian Perez
2026-04-19 20:39:08 +02:00
co-authored by Claude Opus 4.7
parent 0b84e7b7b1
commit f230d9a401
22 changed files with 4134 additions and 0 deletions
@@ -0,0 +1,66 @@
# Phase 9 Verifier V1 — Villa Azul Zod Schema Validation
- Scene: `/tmp/pascal-villa/scenes/a6e7919eacbe.json`
- Scene id: `a6e7919eacbe`
- Scene name: `Villa Azul`
- Declared nodeCount: 56
- Dict size: 56
- AnyNode.safeParse: **56/56 pass, 0 fail**
- parentId integrity: **PASS** (0 issues)
- children[] id integrity (non-site): **PASS** (0 issues)
- SiteNode.children embedded objects (CROSS_CUTTING §2): **PASS** (0 issues)
- Sanity parse (wall/door/window/zone/fence/slab): **PASS**
- **Overall: PASS**
## Per-type counts
| type | total | pass | fail |
| --- | --- | --- | --- |
| building | 1 | 1 | 0 |
| door | 10 | 10 | 0 |
| fence | 5 | 5 | 0 |
| level | 1 | 1 | 0 |
| site | 1 | 1 | 0 |
| slab | 1 | 1 | 0 |
| wall | 12 | 12 | 0 |
| window | 12 | 12 | 0 |
| zone | 13 | 13 | 0 |
| **TOTAL** | **56** | **56** | **0** |
## AnyNode.safeParse failures
_No validation failures._
## parentId reference issues
_None._
## children[] reference issues (non-site nodes)
_None._
## SiteNode.children embedded-object check (CROSS_CUTTING §2)
_None._
Per CROSS_CUTTING §2, `SiteNode.children` is declared as
`z.array(z.discriminatedUnion('type', [BuildingNode, ItemNode]))` and must hold
full embedded building/item objects, not string ids. All other containers use
`string[]`.
## Sanity-parse results (AnyNode.parse, throwing)
| kind | id | status | detail |
| --- | --- | --- | --- |
| wall | `wall_qgrnmxmo0go9yy3q` | PASS | parsed without throw |
| door | `door_333ygsrz65ijnrqv` | PASS | parsed without throw |
| window | `window_8nqg3fvdnb13c0sx` | PASS | parsed without throw |
| zone | `zone_iqi6kkt195pgsdcb` | PASS | parsed without throw |
| fence | `fence_0t2fy6fnsnm5lycx` | PASS | parsed without throw |
| slab | `slab_azul_pool` | PASS | parsed without throw |
## Source
- Script: `packages/mcp/test-reports/villa-azul/v1-schema.ts`
- Input: `/tmp/pascal-villa/scenes/a6e7919eacbe.json`
- Schema: `@pascal-app/core/schema` (AnyNode discriminated union)