Editor: render a layer as one framed row of mark glyphs

A layer is one plotting space with several marks stacked in z-order, so it now renders as a single frame holding its child marks as a row of glyphs, badged with a new 'layers' glyph — instead of offset stacked cards that read as separate spaces and hid the lower glyphs. Each mark stays an individual treeitem, so selection and Alt+up/down z-order reorder are unchanged.

docs/architecture/09-visual-design.md

@@ -377,7 +377,9 @@ drawn on the same 32-grid (stroked where a line reads truer than a fill). It lab
 leaves of the composition wireframe so views read apart at a glance; mark synonyms
 (circle/square → point, trail → line, image → rect) collapse onto it via `markIconName`
 (`mark-icon.ts`), and an unknown or absent mark falls to `mark-generic`. Decorative there
-(the box's `aria-label` names the view), so these are not in the ⭐ icon-only set.
+(the box's `aria-label` names the view), so these are not in the ⭐ icon-only set. The sibling
+`layers` glyph (two offset planes) badges a `layer` node in the same wireframe — a row of marks
+in one frame — as a single shared space rather than a concat's box-per-view.
 
 **Scoped set** — single-surface, glyph **reserved** in the ledger but **not yet in
 the `Icon` registry**:

docs/architecture/10-interaction-and-feedback.md

@@ -228,7 +228,10 @@ schematic of the spec's multi-view composition — `CompositionWireframe`, arch
 tabindex**, arrow keys in **logical (document) order** — Up/Down between nodes, Left → parent,
 Right → first child, Home/End, Enter/Space activate — not spatial, since a mixed horizontal/
 vertical layout makes spatial arrows ambiguous. Each leaf carries a glyph of its mark type (the
-`mark-*` icon sub-family, arch 09 §5) so views read apart at a glance. Selecting a box reveals +
+`mark-*` icon sub-family, arch 09 §5) so views read apart at a glance. A `layer` — one plotting
+space with several marks stacked — renders as **one frame** holding its child marks as a row of
+glyphs, badged as layered (the `layers` glyph), rather than the box-per-view of a concat; each
+mark stays an individual `treeitem` so selection and z-order reorder still work. Selecting a box reveals +
 selects that view's source range in the editor (`AppStore.requestRevealView`) but **does not steal
 focus**, so the wireframe stays the active surface while the editor scrolls to follow; the editor
 selection is the single source of truth. The toolbar glyph appears **only for a composed spec** —

docs/exploration/visual-composition-editing-exploration.md

@@ -149,14 +149,13 @@ DOM `<g name="…">` nodes measurable via `getBoundingClientRect()`. The app alr
 
 ## 6a. Deferred polish (Phase A follow-ons)
 
-Built but parked for a later pass:
-
-- **Mark-type glyph per leaf.** A simplified icon of each unit view's mark (bar/line/point/…)
-  inside its box, so which-is-which reads at a glance without hovering. Needs a small mark→glyph
-  set (its own icon-ledger sub-family).
-- **A more legible `layered` primitive.** Layers currently render as offset stacked rectangles.
-  A dedicated "stacked planes" metaphor — overlapping sheets/disks, like the database glyph's
-  stacked cylinders — would convey "one space, several layers" more elegantly than nudged boxes.
+- **Mark-type glyph per leaf.** _(Shipped.)_ A simplified glyph of each unit view's mark inside
+  its box (the `mark-*` icon sub-family), so which-is-which reads at a glance.
+- **A more legible `layered` primitive.** _(Shipped.)_ A layer renders as **one frame** holding
+  its child marks as a row of glyphs, badged as layered (the `layers` glyph) — not separate boxes,
+  so it reads as one space and stays distinct from a concat (which is box-per-view). The
+  overlapping/stacked-planes options weighed here were rejected: at glyph scale, overlapping
+  line-art muddies the very marks the glyphs exist to show; legibility beat the z-order-depth cue.
 - **Hide the affordance for single-view specs.** _(Shipped.)_ The toolbar glyph appears only when
   the spec has a composition.