Fix four silent drops found by the mobile surface attempt

README.md

@@ -179,8 +179,10 @@ Create a new canvas. If `projectId` is omitted, it lands in the built-in `Untitl
 |-------|------|-------------|
 | `name` | string? | Canvas name |
 | `projectId` | string? | Target project. Defaults to the built-in Untitled project. See `project_list`. |
+| `width` | number? | Artboard width in px, a positive integer (default 1440). A phone screen is `width: 390, height: 844`. |
+| `height` | number? | Artboard height in px, a positive integer (default 900). Each dimension defaults on its own. |
 
-The response also carries a `diversification` signal for the target project: the recently-built structures (newest first) and a hint to differ on at least one taxonomy axis, so successive canvases don't converge on the same layout. It's advisory — never blocking.
+The response also carries the artboard `width`/`height` and a `diversification` signal for the target project: the recently-built structures (newest first) and a hint to differ on at least one taxonomy axis, so successive canvases don't converge on the same layout. It's advisory — never blocking.
 
 ### `canvas_list`
 
@@ -300,7 +302,7 @@ M("nodeId", "newParentId", 0)
 R("nodeId", { type: "text", content: "Replaced" })
 ```
 
-**Returns** `{ ok, nodeIds, results }`. `nodeIds` maps each bound variable to the node ID it created — e.g. `{ "header": "n_a1b2" }` — so you can target those nodes in later calls (bindings only live within a single call). `results` lists each op's outcome in order. If the call wrote a `fontFamily` nothing can serve yet (not cached, registered, or system/generic), a `Font warnings` content item names it — a cache-only check with no network call, so it's a heads-up, not exhaustive.
+**Returns** `{ ok, nodeIds, results }`. `nodeIds` maps each bound variable to the node ID it created — e.g. `{ "header": "n_a1b2" }` — so you can target those nodes in later calls (bindings only live within a single call). `results` lists each op's outcome in order; an op that succeeded but will not fully take effect carries a `warning` (a `shadow` beside an active `shadows`, a chart `highlight` entry outside the series data). If the call wrote a `fontFamily` nothing can serve yet (not cached, registered, or system/generic), a `Font warnings` content item names it — a cache-only check with no network call, so it's a heads-up, not exhaustive.
 
 **Node types:** `frame`, `text`, `rectangle`, `ellipse`, `image`, `icon`, `path`, `component`, `instance`, `toggle`, `checkbox`, `radio`, `select`, `chart`, `skeleton` (loading-placeholder block — token-derived neutral fill; pulses subtly in the live viewer, always static in screenshots/exports so diffs stay deterministic; `pulse: false` opts a block out)
 
@@ -334,7 +336,7 @@ I(panel, { type: "chart", kind: "donut", width: 176, height: 176, centerValue: "
   ] })
 ```
 
-A **bar chart with a selected day** is `highlight: [11]` on the series (selected bar solid, the rest muted; `barGradient: true` fades the muted bars inside the SVG, invisible to the gradient-overuse tell). A **sparkline** is `kind: "sparkline"` — axis-free, latest point emphasized by default, sized for a KPI card corner. `segments[].color` and series `stroke` both take `$chart-1`…`$chart-6` refs.
+A **bar chart with a selected day** is `highlight: [11]` (or `highlight: 11`; `null` clears it) on the series (selected bar solid, the rest muted; `barGradient: true` fades the muted bars inside the SVG, invisible to the gradient-overuse tell). An entry that isn't an index within the series `data` emphasizes nothing, and the `batch_design` op result warns about it. A **sparkline** is `kind: "sparkline"` — axis-free, latest point emphasized by default, sized for a KPI card corner. `segments[].color` and series `stroke` both take `$chart-1`…`$chart-6` refs.
 
 Multi-series in one node; x positions are data indexes (a shorter series stops early against a longer one — booked months vs a full-year target); `xDomain`/`yDomain` default from the data (bars floor at 0); `kind: "bar"` renders grouped bars from the same series model. Dash the projected series, solid the actuals. Editing one value is a one-prop edit — never hand-compute path `d` strings for a chart.
 
@@ -406,8 +408,8 @@ Render canvas to PNG (returned as base64 image).
 |-------|------|-------------|
 | `canvasId` | string | Canvas ID |
 | `nodeId` | string? | Specific node to capture |
-| `width` | number? | Viewport width (default 1440) |
-| `height` | number? | Viewport height (default 900) |
+| `width` | number? | Viewport width (default: the canvas's artboard width) |
+| `height` | number? | Viewport height (default: the canvas's artboard height) |
 | `fullPage` | boolean? | Capture the whole design instead of one viewport — default false, so existing captures (diffs, responsive sets, the pattern gate) are byte-identical. A canvas taller than its artboard otherwise cuts off at the artboard height. `export` takes the same option (ignored for PDF, which paginates) |
 | `scale` | number? | Device scale (default 2) |
 | `theme` | string? | `"dark"` renders the design system's dark token layer (`dark.colors`/`dark.elevation` overrides); default light — a no-op without a dark layer |
@@ -567,8 +569,8 @@ Export a canvas or specific nodes to files on disk.
 | `format` | string | `"png"`, `"jpeg"`, `"webp"`, or `"pdf"` |
 | `outputPath` | string | Directory to save files |
 | `nodeIds` | string[]? | Specific nodes to export (default: full canvas) |
-| `width` | number? | Viewport width (default 1440) |
-| `height` | number? | Viewport height (default 900) |
+| `width` | number? | Viewport width (default: the canvas's artboard width) |
+| `height` | number? | Viewport height (default: the canvas's artboard height) |
 | `scale` | number? | Device scale (default 2) |
 | `theme` | string? | `"dark"` renders the design system's dark token layer (`dark.colors`/`dark.elevation` overrides); default light — a no-op without a dark layer |
 
@@ -765,8 +767,8 @@ Compare two canvases visually. Returns a diff image with changed regions highlig
 |-------|------|-------------|
 | `canvasId1` | string | First canvas ID |
 | `canvasId2` | string | Second canvas ID |
-| `width` | number? | Viewport width (default 1440) |
-| `height` | number? | Viewport height (default 900) |
+| `width` | number? | Viewport width (default: the first canvas's artboard width) |
+| `height` | number? | Viewport height (default: the first canvas's artboard height) |
 | `scale` | number? | Device scale (default 1) |
 
 ### `get_feedback` / `resolve_feedback`
@@ -799,26 +801,28 @@ Returns `{ canvasId, genre, previous, relaxes, note?, versionHashUnchanged }` 
 
 Auto-score a design against quality heuristics. Returns an overall score (0–100), per-category scores, per-node actionable issues, and a **`directive`** — a present/keep-working verdict. Designed for generator-evaluator loops: build with `batch_design`, score with `canvas_evaluate`, fix the issues targeting the returned `nodeId`s, repeat until the directive says `READY`.
 
-The **`directive`** is the operating contract that keeps unfinished work off the user's screen. Readiness turns on **blocking findings, not the score**: it reads `READY TO PRESENT` when there are zero errors, zero warnings, **and** zero cliché tells; otherwise `NOT READY` with what to resolve. (Cliché tells block even at info severity — they're the slop signal; pure advisories like "consider extracting components" are optional.) The score is always reported alongside the directive and remains the quality bar to aim for (**> 95**) — but it no longer withholds readiness on its own: once nothing blocking is left, a design at or below the bar still reads `READY`, and the directive says so explicitly, naming the advisory findings as the remaining lever. Resolve everything blocking and re-evaluate until it says `READY` before showing the design to the user.
+The **`directive`** is the operating contract that keeps unfinished work off the user's screen. Readiness turns on **blocking findings, not the score**: it reads `READY TO PRESENT` when there are zero errors, zero warnings, **and** zero cliché tells (and no `FRAGILE` stress verdict, see below); otherwise `NOT READY` with what to resolve. (Cliché tells block even at info severity — they're the slop signal; pure advisories like "consider extracting components" are optional.) The score is always reported alongside the directive and remains the quality bar to aim for (**> 95**) — but it no longer withholds readiness on its own: once nothing blocking is left, a design at or below the bar still reads `READY`, and the directive says so explicitly, naming the advisory findings as the remaining lever. Resolve everything blocking and re-evaluate until it says `READY` before showing the design to the user.
 
 Open point-and-tell comments block too: when the canvas has any, the result carries `openFeedback: n` and the directive stays blocking even at a READY score — the human's note outranks the heuristics. Read them with `get_feedback`, address each, close them with `resolve_feedback`.
 
+A `FRAGILE` stress run blocks too. Once `canvas_stress` has recorded a verdict for the canvas, the result carries **`stress`**: `{ verdict, warnings, current }`, where `current` says whether that run was on the canvas's present `versionHash`. A `FRAGILE` verdict keeps the directive `NOT READY` even with zero evaluate issues, until a full `CLEAN` run (all five perturbations) replaces it; a subset run can record `FRAGILE` but never clear it. When `current` is false the directive tells you to re-run stress to confirm your fix. The verdict lives in server memory, so a restart forgets it. `canvas_evaluate` never runs stress itself.
+
 Whenever the `cliche` category runs, the result also carries a **`genre`** field auditing the genre decision: `{ active, source, relaxed, notRelaxed }` — the genre in effect, whether it came from the `genre` param or the canvas's provenance stamp, the tells it relaxed, and `notRelaxed` (`[{ tell, relaxedBy }]`) — tells still flagging that a *different* genre would relax. A score pinned by tells listed in `notRelaxed` usually means the wrong genre was declared.
 
 Whenever the `coverage` category runs on a **base** canvas, the result also carries a **`coverage`** field: `{ dataBearing, states, missing }` — whether the screen carries data-bearing content, which state variants (`canvas_add_variant`) already exist, and which demanded states are missing. It's absent on variant canvases (a variant *is* a designed state, so it doesn't demand states of itself).
 
 | Param | Type | Description |
 |-------|------|-------------|
 | `canvasId` | string | Canvas ID to evaluate |
-| `mode` | `"fast"` \| `"detailed"` \| `"llm"` | `"fast"` = JSON-tree analysis only (<100ms). `"detailed"` adds Puppeteer-based pixel-level overlap checks. `"llm"` runs fast-mode heuristics plus a vision-model critique (provider picked from `FRAMESMITH_LLM_PROVIDER` or whichever of `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` is set — costs one paid API call per invocation). Default `"fast"`. |
+| `mode` | `"fast"` \| `"detailed"` \| `"llm"` | `"fast"` = JSON-tree analysis only (<100ms). `"detailed"` adds Puppeteer-based pixel-level overlap checks, laid out at the canvas's artboard size. `"llm"` runs fast-mode heuristics plus a vision-model critique (provider picked from `FRAMESMITH_LLM_PROVIDER` or whichever of `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` is set — costs one paid API call per invocation). Default `"fast"`. |
 | `categories` | string[]? | Subset of `spacing`, `color`, `typography`, `structure`, `consistency`, `cliche`, `coverage`, `usability`. Defaults to all. |
 | `genre` | string? | Style that relaxes specific `cliche` gates — `"material"` allows purple and white elevated surfaces; `"dashboard"` (alias `"data"`) allows realistic figures on data-dense screens; `"commerce"` (alias `"checkout"`) allows the money on a transactional screen — cart, checkout, order confirmation, billing history. Defaults to the canvas's provenance preset if stamped (`canvas_set_genre` stamps it durably, no token churn). Pick by what the screen is *for* (read screens with published figures → `dashboard`; a screen you complete a purchase on → `commerce`; editors/admin forms → `material`; a pricing page is marketing and gets none of them), not what it contains — the result's `genre` field audits the choice. |
 
 **Categories and what they check**
 
 | Category | Weight | Checks |
 |----------|--------|--------|
-| `spacing` | 20 | Off-scale padding/gap values, plus spacing variety measured against the canvas's **declared** scale (`set_variables`/`generate_scale` spacing tokens) — only values *not* on that scale count toward sprawl, so a design built entirely from its own multi-step scale (framesmith's generated scale ships nine: `space-3xs`…`space-3xl`) isn't penalized for using it. No declared scale falls back to the old total-unique-values count, with a suggestion to declare one |
+| `spacing` | 20 | Off-scale padding/gap values (`gap: 0`, a flush stack, is never flagged), plus spacing variety measured against the canvas's **declared** scale (`set_variables`/`generate_scale` spacing tokens) — only values *not* on that scale count toward sprawl, so a design built entirely from its own multi-step scale (framesmith's generated scale ships nine: `space-3xs`…`space-3xl`) isn't penalized for using it. No declared scale falls back to the old total-unique-values count, with a suggestion to declare one |
 | `color` | 25 | WCAG AA contrast ratios for text against nearest background — checked in BOTH themes when a `dark` token layer exists (dark-run issues carry `theme: "dark"`, score takes the worse theme, and dark failures point at the dark layer instead of carrying a literal fix). WCAG-passing pairs that are perceptually weak by **APCA** get an info advisory (Lc vs ~75 body / ~60 large) — APCA is a candidate method, not a standard; it never blocks |
 | `typography` | 20 | Type-scale ratios (1.1–2.0; adjacent sizes both declared as typography tokens are pinned and skip the check), font-family count, weight variation — plus score-neutral craft advisories: measure (prose past ~75 characters per line; warning past 90), tracking (display text ≥28px without negative letter-spacing, autofixable), and tabular numerals (numeric table columns without `tabularNums`, autofixable) |
 | `structure` | 15 | Tree depth, naming coverage, design-token usage %, component reuse, unknown icon names (an icon that resolves in neither set renders an empty slot: a directive-blocking warning, −10 structure score each; autofixable when one spelling resolves) |
@@ -936,7 +940,7 @@ Apply the ops by joining them with newlines and passing to `batch_design`, then
 
 ### `canvas_stress`
 
-Content stress test — does the design survive real data? Re-renders the canvas under hostile-but-realistic content perturbations and reports exactly what broke, by node id. Never mutates the canvas.
+Content stress test — does the design survive real data? Re-renders the canvas under hostile-but-realistic content perturbations and reports exactly what broke, by node id. Never mutates the canvas. The verdict is remembered for `canvas_evaluate`: any run with warnings records `FRAGILE` (its directive then stays `NOT READY`), and only a run of the full default set records `CLEAN`, so a passing subset never clears an earlier `FRAGILE`.
 
 | Perturbation | What it does |
 |---|---|