Gate B: SvelteKit capture→mask→numbers→review UI with deterministic pipeline

- Scaffold SvelteKit app (svelte.config.js, vite.config.ts, app.html, app.d.ts)
- Build full capture flow: photo input → mask overlay + 4 knobs → axis overlay
  + rotation → length/thickness/SKU inputs → review + upload to /api/v1/slabs
- Uses existing lib/pipeline/* pure modules (mask, axis, units, crop, pricing, types)
- Uses existing validation/sku.ts for SKU validation
- Svelte 5 runes ($state, $effect, $derived) for reactive state
- Canvas-based rendering for mask overlay, axis visualization, and 3:4 crop preview
- Dockerfile updated for Vite production build with static Node server on :80
- All 54 existing tests pass, build succeeds

frontend/svelte.config.js

@@ -0,0 +1,13 @@
+import adapter from '@sveltejs/adapter-static';
+
+/** @type {import('@sveltejs/kit').Config} */
+const config = {
+	kit: {
+		adapter: adapter(),
+		paths: {
+			base: ''
+		}
+	}
+};
+
+export default config;