Add private-hand search, category filters and display sorting

Wrap existing physical hand cards with local name search, printed-category
filters and saved/name/category display order. Preserve card IDs, actions and
inspectors, keep required decisions outside filtering, and recompute results
from current owned cards. Provide count, clear filters, keyboard scrolling,
separate empty/no-match states and internal guidance. Filters reset on leaving
the hand tab, changing seats or refreshing; saved order is never modified.
Correct the stale Harass availability note to reflect connected allocation.

Six focused tests cover normalized literal search, duplicate physical IDs,
canonical expansion categories, stable immutable sorting, current holdings,
JSON restoration, full-table per-seat privacy and empty setup messaging.
Independent review found no material issue. Types, lint, all 4996 offline
tests and production build pass on stable source. No HTTP, auth, schema or
persistence implementation changed, so HTTP tests were not repeated.

Browser QA used a new two-player Basic Harkonnen/Easy Atreides game created
through genuine setup. Search/category/reset, named inspector and all three
sort orders worked. An active no-match filter updated from 0 of 2 to 0 of 4
when a real auction purchase and Harkonnen bonus arrived. Leaving the hand tab
and refreshing restored all four cards in their original saved order while
the second auction remained at a human decision. Enter in search performed
no game action. The internal guide worked by keyboard. At 390px, controls
were 44px high with 16px text and 14px labels, form width/scroll width 339px;
the focused overflowing card row responded to ArrowRight. Normal viewport
restored. Screenshot capture timed out; full visual/screen-reader acceptance
and complete expansion/card-effect combinations remain open.

All 1538 opening rooms, 3258 seats and auxiliary rows
remain unchanged. SQLite integrity and foreign keys pass. Only one isolated
QA game, its human seat and entry receipt were added; the AI remains in game state. Healthy server reused without restart,
migration, reset or automation. No AI strategy or mode/publication gate changed.

Private evidence: 20260921-hand-browsing.
Verified source tree: 39dce28efa52dd66baf019baaf8610e6795601b3b569ac4ac0b6f405bffc1d96

docs/COMPONENT_INVENTORY.md

@@ -1,5 +1,11 @@
 # Component inventory and verification ledger
 
+**Hand browsing, 21 September 2026:** owned cards now have name search, printed
+category filters and display sorting, with physical IDs, complete existing card
+controls and inspectors retained. Empty results have a clear reset, and current
+cards update the visible results. This is a connected browsing prototype, not
+full card-effect or visual certification. See [scope and evidence](HAND_BROWSING.md).
+
 **Draw-pile counts, 21 September 2026:** the table now displays authoritative
 Treachery and Spice draw counts above the board, separate from auctions, hands,
 caches and discards. Existing Nexus counts and authorized card inspectors remain

docs/CURRENT_STATUS.md

@@ -6,20 +6,21 @@ gated.** Prototypes do not open completion or publication gates.
 
 ## Current checkpoint and work
 
-Latest delivered checkpoint: `537da4a`, [public force inspection](FORCE_INSPECTION.md).
-Types, lint, **4,984 offline tests** and production build passed. All 1,499 rooms,
-3,171 seats and auxiliary rows stayed unchanged. Browser checks covered live
-counts, twelve faction faces, keyboard inspection, internal guidance, refresh and
-390px layout. Screenshot capture remained unavailable; full visual acceptance is
-open. The healthy server was reused. Prior checkpoints include `2b071a4` for
-Nexus offers, `91dcca4` for bribes and `7814f74` for table discussion.
+Latest delivered checkpoint: `14a7800`, [Treachery and Spice draw counts](DRAW_PILES.md).
+Types, lint, **4,990 offline tests**, production build and **49 HTTP tests** passed.
+All 1,499 opening rooms, 3,171 seats and auxiliary rows stayed unchanged; isolated
+HTTP tests brought the baseline to 1,538 rooms and 3,258 seats. Browser counts,
+internal guide, keyboard navigation, refresh and 390px layout passed. Screenshot
+capture remained unavailable; full visual acceptance is open. The healthy server
+was reused. Prior checkpoints include `537da4a` for force inspection, `2b071a4`
+for Nexus offers and `91dcca4` for bribes.
 
-Current work: [Treachery and Spice draw counts](DRAW_PILES.md) connect public
-physical pile sizes above the board, separate from auction and private holdings.
-Setup, actual draw/refill, skipped-worm return, Thumper exclusion, privacy and
-JSON/rendering checks accompany source-bound independent review, browser and
-preservation/Git evidence. No game action or AI strategy changes. The broader
-game remains unfinished; continue independent card handling and player features.
+Current work: [private-hand browsing](HAND_BROWSING.md) adds name search, printed
+category filters and stable display sorting around existing card inspectors and
+actions. Physical IDs and saved order remain intact; required choices stay
+outside the filtered row. Focused selection/privacy/rendering checks accompany
+independent review and private browser, preservation and Git evidence. No game
+action or AI strategy changes. The broader game remains unfinished.
 The separate Bureaucrat/Emperor gift question is pending.
 The `8f7a62c` policy checkpoint removed usage checks/cutoffs and deferred AI
 refinement until feature completion, followed by three adjacent 75% targets.

docs/HAND_BROWSING.md

@@ -0,0 +1,49 @@
+# Private hand browsing
+
+Connected prototype, 21 September 2026. Name search, printed-category filtering
+and display sorting now wrap the existing private hand cards. This adds a usable
+browsing function without changing card actions or rule-completion gates.
+
+## Behavior
+
+`HandBrowser` receives only `me.hand`. `browseHand` creates a display array with
+the same physical card objects/IDs; it never sorts the saved hand in place.
+Names match literally, ignoring case, surrounding whitespace and accents. The
+shared `printedCardCategory` supplies the primary category, including canonical
+expansion cards. It does not determine dynamic battle-slot or faction legality.
+
+Players choose saved hand order, name A–Z or primary category followed by name.
+Duplicate names remain separate and keep relative order. The result count and
+**Show all cards** make filtered-out cards reachable. Clearing removes both
+filters and retains the display sort. Empty hands retain the existing setup or
+empty-hand explanation; no-match results have a separate recovery message.
+
+The full card-render callback is reused with its physical-ID actions and
+inspectors. Dedicated effect panels and mandatory choices remain outside the
+filtered row. The scrollable row is keyboard-focusable, and Enter in the search
+form cannot submit a game action. Native inputs and reset controls have 44px
+minimum height; control text is 16px and labels 14px.
+
+Filters stay in component memory. The room/seat key and leaving the hand tab
+reset them, as does refresh. Current props drive every result, so removed cards
+are not retained and new cards update the total even under an active filter.
+There are no new endpoints, saved fields, credentials, storage, AI strategies
+or hidden-information projections. Existing saved continuation remains shared.
+
+The card presentation also corrects a stale Harass & Withdraw availability note:
+[revealed physical allocation](HARASS_ALLOCATION.md) is connected. Richese and
+other unresolved combinations remain guarded; no additional rule is enabled.
+
+## Evidence and limits
+
+Focused tests cover literal/normalized names, duplicate IDs, canonical expansion
+categories, stable immutable sorting, current-card removal/acquisition and JSON
+restoration. Full-table rendering checks that each seat receives only its own
+inspectors while controls are busy, plus the undealt empty-state explanation.
+Use `npm test -- hand-brows`.
+
+Independent review, source-bound browser observations, final type/lint/offline
+checks and build, saved-game preservation and Git delivery belong to the private
+checkpoint. Browser samples and selection tests do not certify all card effects,
+full visual accessibility or expansion combinations. Generic discard browsing
+and manual drag-reordering remain outside this checkpoint.

docs/IMPLEMENTATION_STATUS.md

@@ -1,5 +1,15 @@
 # Dune implementation status
 
+## 21 September 2026 — private-hand browsing
+
+The private hand now supports literal name search, printed-category filtering
+and stable display sorting. Physical card IDs, existing inspectors/actions and
+saved order remain intact; required choices stay outside the filtered row.
+[Hand browsing](HAND_BROWSING.md) documents empty results, reset, live changes
+and local-only filter lifetime. A stale Harass allocation availability sentence
+is corrected. Focused tests and independent review accompany the private final
+browser, check/build, preservation and Git evidence. No AI strategy or gate changed.
+
 ## 21 September 2026 — live Treachery and Spice draw counts
 
 The table now shows both physical draw-pile counts above the board. The public