Gold Standard row/expand contract

The row and expand contract every widget in the registry follows.

Every REGISTRY widget renders its content as N≥1 rows, using Skew.tsx as the reference implementation. A widget's data shape (chart, table, single app-scoped scalar) is never by itself a reason to skip this: the four named exceptions below (one histogram toggle, three partial exceptions) are the only ones. Any further exception must get its own entry here, justified in the same PR that introduces it.

• Tier A (universal): WidgetCard chrome, an impact-band left-border + dot, ALL-CAPS TagBadges, a SortModeToggle when canToggleSort is true, a finding-anchor ref on any focusable/deep-linkable unit, and an early null return on an empty catalog filter (except Core Usage by Locality, the one always-mounted widget that renders unconditionally from appModel via alwaysMountedWidgets()/isAlwaysMountedType() rather than early-returning on an empty catalog filter).
• Tier B (the row/expand pattern): a row's collapsed state shows its core metric(s), ImpactEstimate, any config-hint code snippet/list, and docs links, all unconditionally. The fix is stated once per card: WidgetCard's fixFor prints each distinct coreFindingGenericRecommendation above the body, so rows don't repeat a recommendation that restates their number. Every finding-row widget passes it, Shuffle I/O and the Skew, Stage Shape and Tiny Tasks groups (StageFindingGroup) included; Incomplete Run is the one that doesn't. Where the detector's sentence is a row's only measurement (Excessive Small Files, Missed Broadcast Join, Oversized Broadcast Join, Config Audit), the row shows its measured half, split off by core's recommendationParts. Three rows keep their own sentence because their type has no generic fix for them: Incomplete Run, and a data-unavailable Memory Utilization or Cache Storage row, which names the setting to enable. Confidence and evidence are unconditional too: RowStatusCluster (src/view/RowStatusCluster.tsx) is a single, fixed-position pill combining both. Every widget that carries one wraps it in AdvancedOnly (src/view/AdvancedOnly.tsx), so it only renders at the Advanced density tier: this is a content-visibility gate (Basic vs. Advanced), not a per-row collapse, and it's applied consistently across every adopting widget. No widget gates confidence or evidence behind a per-row click. A widget-header RowStatusCluster (passed as WidgetCard's statusBadge prop) carries one further, unrelated gate on top of AdvancedOnly: open && statusBadge (WidgetCard.tsx's header) keeps it out of the collapsed header, matching the comment on statusBadge itself ("kept separate so tag/impact badges stay visible in the collapsed summary while a RowStatusCluster-style ... marker doesn't"). This is the card's own expand/collapse, not a separate reveal-on-click built for the cluster, and it composes with density as an AND: a header cluster needs both Advanced tier and an expanded card. A row with neither renders no cluster at all (RowStatusCluster returns null). Confidence renders as plain, non-interactive text ("{confidence} confidence", no further suffix) with a title + aria-describedby tooltip carrying the full validationRequired detail: never itself a click target. Evidence renders as the cluster's one real click target: a button whose visible text is just the evidence label (e.g. "SQL plan") but whose accessible name always carries the "Evidence: ..." prefix via an explicit aria-label, regardless of what else sits nearby. Clicking it calls revealEvidence. RowStatusCluster's fixed slot is the row's own stage-pill/title line (justify-between, cluster right-aligned), the same position in every adopting widget, not floating with the row's detail below. Lists longer than VISIBLE_LIMIT (6, packages/core/src/format-utils.ts) get page-based navigation (Previous/Next, usePagedRows/RowPagination, src/view/usePagedRows.ts/src/view/RowPagination.tsx) instead of rendering unconditionally: the one sanctioned cap mechanism across every Tier B widget. A Tier B widget whose rows are also triage-routable (wires useFindingAnchor) must pass its routeIndex (the routed finding's position in the paginated list) into usePagedRows, or a deep-linked route to a finding beyond the first page silently degrades to focusing the widget's disclosure title instead of the row.

The following widgets carry a RowStatusCluster, all Advanced-only: Spill.tsx (confidence only), ConfigAudit.tsx's widget-header control (evidence only, a widget-wide constant rather than per-finding data, since the evidence key doesn't vary per row; also the control shown in the widget's empty-findings/no-data states), MemoryUtilization.tsx's widget header (confidence + evidence, one marker, since executorMetrics is widget-wide; ExecutorUtilization.tsx's rows carry neither), each of the six Plan Advisor widgets' (DuplicatePlanSubtree.tsx, SmallFiles.tsx, NestedLoopJoin.tsx, PythonUdf.tsx, UnderBroadcast.tsx, OverBroadcast.tsx) widget header (confidence + sqlPlan evidence, one marker per widget since each is its own finding type; no per-row control), ScalingSim.tsx (two call sites, evidence only), CoreUsageArea.tsx, EfficiencyModel.tsx and PlanView.tsx (three call sites) (these last confidence-only; all standalone rather than per-finding-row: a card-header badge, a widget-level single marker, a plan-tree node/summary-row marker, or an "unavailable data" message: the same component, same visual language, regardless of where it sits); and, because skew/straggler/gc disclose their own unvalidated noise-floor thresholds (confidence only, per-row, no evidenceKey passed), GcPressure.tsx's rows, Straggler.tsx's rows, CachingOpportunity.tsx's rows (the cachingOpportunity detector scales confidence per finding via cachingReuseConfidence, so the badge sits next to each row rather than as a single caveat below the table), and the shared StageFindingGroup.tsx row (adopted by Skew.tsx/StageShape.tsx/TinyTask.tsx, though only skew findings carry a confidence field).

One named exception:

• Skew.tsx/StageShape.tsx/TinyTask.tsx: their per-row histogram toggle (ExpandToggleButton + useExpandableRow) gates a lazily-fetched duration histogram, unrelated to confidence/evidence. It sits alongside RowStatusCluster in the shared StageFindingGroup.tsx row; the two controls coexist per row. ExpandToggleButton is single-purpose (always the task-detail toggle) since these three are its only adopters.

No toggle at all, unconditional content, since these widgets carry no confidence/evidence display: IncompleteRun.tsx, ShuffleIO.tsx, PartitionSizing.tsx, StageFailed.tsx, TaskFailures.tsx, RetryWaste.tsx, ColdStart.tsx, SlowHost.tsx, StageSlowness.tsx, SpeculationWaste.tsx, ExecutorCountChart.tsx, ExecutorUtilization.tsx, JobFailures.tsx, CacheUtilization.tsx and AutoscalingChurn.tsx.

Three named exceptions to specific pieces of the contract, not to the row wrapper or pagination, which still apply to all three:

• CoreUsageArea.tsx's non-local-stage list is a derived stat breakdown (computeCoreLocalityRatio's topStages), not a Finding[]: there is no per-row finding or recommendation to reveal. Paginated like every other Tier B list, but with no per-row expand.
• CacheUtilization.tsx's RDD <Table> holds reference columns (name, storage level, partitions, memory, disk bytes) with no recommendation attached to any row. Paginated like every other Tier B list, but with no per-row expand. Its separate finding list below the table is full Tier B (unconditional, no toggle, per the paragraph above).
• IncompleteRun.tsx leads with no metric of its own: this app-scoped, single-finding widget has no natural lead figure distinct from its title/badge, unlike every other Gold Standard widget, which leads with a percentage, duration, or count.