Module model 

3D model-viewer embed synthesizer.

Receives a TitleParams (Stage 2 dispatcher already parsed it), the source URL, and an AssetSnapshot. Emits final <model-viewer> HTML — preserving the shape moss-core’s ModelViewerRenderer used to emit before Phase 0’s Stage 1 migration.

Source byte shape (pre-Phase-0, see commit 689d975e9^, crates/moss-core/src/resolve/embed_renderer.rs:723):

<model-viewer class="moss-embed" data-type="3d"{data-width} src="{src}"
  camera-controls auto-rotate touch-action="pan-y" loading="lazy"{style}>
</model-viewer>

{data-width} is data-width="VALUE" (with leading space) when the wrapper width is set (body | wide | page | screen), empty otherwise. {style} is style="width:W" or style="width:W;height:H" when sizing params are present, empty otherwise.

Stage 2 reads the boolean flags (camera-controls, auto-rotate, ar) from TitleParams — when the wikilink grammar surfaces them as param=true they appear as bare attributes on the element.

Functions

synthesize_model_html

Synthesize 3D model-viewer embed HTML for Tag::Link with moss:kind=3d title.