Replace addons with conventional handles

Provide the single-loadout Scene API, SAB-backed meshes, cameras, materials, lights, workers, post effects, and tutorial playgrounds. Batch matching row growth so active GPU loadouts refresh once.

Amp-Thread-ID: https://ampcode.com/threads/T-01a01380-b478-77d0-84a0-102880a5c5ae
Co-authored-by: Heaust Azure <heaust.azure@gmail.com>

docs/guide/scene-and-sab.md

@@ -0,0 +1,44 @@
+# Scene and shared data
+
+Think of every handle as an array index, not an object mirrored into core. `Node.position`, `Node.quaternion`, and `Node.scale` are views into separate flat SOA arrays.
+
+## Direct transform movement
+
+```ts
+import { Node } from "@yawn/handles";
+
+const pivot = new Node(scene, { position: [0, 1, 0] });
+await pivot.ready;
+
+canvas.addEventListener("pointermove", (event) => {
+  pivot.position[0] += event.movementX * 0.002;
+  pivot.position[1] -= event.movementY * 0.002;
+});
+```
+
+The pointer handler sends no messages. The typed-array view points directly into the arena shared with the render worker. The camera helpers use this same pattern; see [Cameras and controls](/guide/cameras).
+
+## Add an application-specific row
+
+```ts
+const particles = await scene.ensureRows("particleVelocity", 10_000, 16, "f32");
+particles.row(42).set([1, 0, 0, 0]);
+```
+
+Rows are 16-byte-stride-aligned and arena allocations are 64-byte aligned. Formats are `f32`, `u32`, or `i32`.
+
+## Timing and render skipping
+
+Core always creates `info` as:
+
+```text
+[deltaTime, frameCount, elapsedTime, targetFps, skipRender, 0, 0, 0]
+```
+
+```ts
+const info = scene.array("info").row(0);
+info[4] = 1; // keep timing, skip GPU work
+info[4] = 0; // resume rendering
+```
+
+Use messages for rare control changes (`setFps`, graph updates, allocation); use SAB writes for existing hot state.