feat(simulation): add Gaussian UGV runtime pipeline

docs/20_SIMULATION_WORLD_SURFACE_BRIEF.md

@@ -1,6 +1,7 @@
 # Simulation Worlds — product-surface brief
 
-Status: accepted for the first working vertical slice on 2026-08-26.
+Status: accepted for the first working vertical slice on 2026-08-26; visual/physical stage boundary
+revised on 2026-08-27.
 
 ## Job story
 
@@ -57,8 +58,12 @@ Mission Core backend
 DC Gaussian Pipeline
   ├─ TUS bundle or archive admission
   ├─ secure ZIP/RAR/7z normalization
-  ├─ native Vulkan SplatTransform build on Worker 006
-  └─ bounded collision simplification + mandatory Draco publication
+  └─ native Vulkan SplatTransform visual build on Worker 006
+
+Optional physical-mesh pipeline (separate job; experimental)
+  ├─ source-mesh discovery or explicit mesh generation
+  ├─ physics-oriented cleanup and geometry budget
+  └─ compressed, digest-bound publication
 ```
 
 The browser never receives the Worker token. Mission Core does not embed archive-format behavior
@@ -76,13 +81,18 @@ SplatTransform GPU command through a confined filesystem spool to the pinned nat
 on the RTX 4090. It does not install into or share the Python, CUDA, Triton or computer-vision
 environments on the host.
 
-Collision publication is deliberately two-stage. SplatTransform generates smooth collision and
-SVO query data at a scene-type voxel size (`0.2 m` outdoors, `0.1 m` indoors/objects). The isolated
-pipeline container then bounds the debug/physics mesh to a default 250,000-triangle target and
-always writes `KHR_draco_mesh_compression` before the artifact digest is sealed. Simplification
-controls decoded geometry cost; Draco controls transfer and stored bytes and is never treated as a
-replacement for a physics mesh budget. The browser loads the result through a pinned local Draco
-WASM decoder with no external decoder request.
+Project processing is split at a durable product boundary. The mandatory first job builds only the
+preview and streamed Gaussian assets required for visual inspection. It never generates collision
+geometry, so a location can reach `ready` without paying the time, GPU-memory and storage cost of a
+physical mesh.
+
+Physical geometry is an optional second job started only after the visual world is ready. Its first
+candidate source is a mesh already present in the uploaded export (for example a PLY in
+`Mesh_Files`); generation from the Gaussian cloud is a fallback experiment, not the default path.
+The second-stage contract, cleanup method and acceptance gates are intentionally separate from the
+visual build. When that stage publishes a GLB, simplification still controls decoded physics cost
+and Draco controls transfer/storage bytes; compression is not treated as a replacement for a
+physics mesh budget.
 
 Visual and collision layers retain independent X/Y/Z correction settings, while PlayCanvas world,
 camera, navigation and future physics stay in the canonical Y-up coordinate system. Quality, both
@@ -105,7 +115,9 @@ Primary implementation references:
 - encrypted, linked, traversing, duplicate and over-limit archive entries fail closed;
 - project status is durable and reflects provider state without fabricated percentages;
 - ready artifacts are imported digest-bound and served from Mission Core same-origin URLs;
-- every published collision GLB is simplified to the configured budget and requires Draco;
+- the mandatory build requests preview and streamed Gaussian outputs with collision disabled;
+- a visual project reaches ready state without a collision artifact;
+- physical mesh preparation is a separate explicit job and never blocks visual inspection;
 - edit changes project metadata; delete removes both the Mission Core project and terminal provider
   job;
 - a ready project mounts direct PlayCanvas Engine and loads Streamed SOG with preview fallback;