KTools 4.0

Verified locally 2026-10-07. On the prepper station select My Shift (index.php?view=my_shift). Layout updated locally 2026-10-07: My Shift is immediately after Add Record in the prepper navigation. Navigation order updated locally 2026-10-07: Add Record, My Shift, Prep List, Search, View, Edit, Case Pack, Misfits, Damage Out, Tasks, Pick List, Supply, Days Off, Labels, Reports, Message, Manual, Scanner Test, then Legends when enabled. Links retain their existing destinations and active-page highlighting. Its summary tiles, section headers, task links, Refresh action and recent-work table now follow the shared station theme, muted guidance and card styling. On phones the summary tiles use two columns and task cards stack. Work totals, identity, approvals and game behavior are unchanged. A remembered, active prepper name is required. The page shows that prepper's saved units, records and different SKUs for the warehouse-local day, latest 20 submissions, all-date pending mission reports and reported pallet counts, open Productivity SKUs/Supply Runs, last 10 reports and existing game level/XP/coins. Refresh reloads the snapshot. Tasks remain available when Legends is off; retained progress is labeled. Dates on submissions are UTC. Totals use original record creation time and exclude deleted records; units are quantities, not scan attempts. Unknown names create no profile. Viewing never submits work or grants rewards. Use View tasks to report missing SKUs/restocked supplies through existing approval/PIN rules, or Open Warehouse Legends for game actions. If unavailable, refresh and ask a supervisor to check database updates.

Verified locally 2026-10-07. Select Scanner Test (index.php?view=scanner_test), then Start camera test and allow browser camera access. Use the shared Camera selector to try each lens. Hold a clear supported UPC/EAN, Code 128 or Code 39 label steady in good light. Stop camera records session duration, successful reads and first-read delay; switching lenses also records the prior session. Latest 20 test results remain only on this page. Repeated reads of a held label may count again after 1.2 seconds; these are decoder successes, not distinct items. For hardware scanners, click Hardware scanner test and scan. Enter/Tab are captured without submitting; missing suffix appears after a pause. Clear test empties the input. Camera closes when the page is hidden or left. Camera images/barcode values are never saved; existing Device Usage stores only technical timing/counts. The last selected camera remains the station preference. No inventory or game action occurs. Use HTTPS, clean the lens, close other camera apps, check site permissions, and ask an administrator to review Device Usage if scanning fails. Physical-device testing remains pending.

Verified locally 2026-10-07. Admin Tools → System Health (admin/system-health.php) requires admin_tools.health.view, granted to admins by seed 0074; grant others through Role Manager. Each tool additionally requires its own view permission. Refresh checks reads database connectivity, unapplied updates, release file comparison, Iron Bot heartbeat, printer heartbeat/failed jobs and device problems. Healthy, Needs attention, Information and Unknown states include next steps and permitted review links. A missing heartbeat does not prove broken hardware or guarantee a model answer. Device counts describe cumulative occurrences in groups last seen today in the account time zone, not exact newly occurring events; external catalog thumbnails are information. Failed checks show Unknown and leave the other cards available. This page never retries saves, prints jobs, applies migrations or repairs files.

Redesigned page (2026-10-07): Admin Tools → System Health (permission admin_tools.health.view) now opens with an overall status, then What needs attention: one tile per tool you can view (release files, database updates, Iron Bot, print stations and jobs, device problems, save failures) with the next step and a link. Below are quick metrics, monitored subsystems and server environment details (PHP and database versions, memory and upload limits). Everything is read-only; nothing is repaired automatically.

Verified locally 2026-10-07. Admin Tools → Operational Reports (admin/operational-reports.php): choose a permitted report and View report. The page and Iron Bot report access require reports.operations.view (seed 0075, inherited by roles already viewing a supported source) plus the selected source permission. Units by prepper today requires reports.custom.view; its account-local start is inclusive and next midnight exclusive, using saved record creation time and excluding deletions. Printers without a recent heartbeat requires labels.station.run and uses the shared 30-second threshold for active printers only. Mission approvals requires legends.command.view; kiosk approvals legends.rewards.view; supply requests supply.requests.view. Approval reports are current all-date status, not historical status on a chosen day, and reported pallets are not approved rewards. Iron Bot can use these fixed SELECT reports from any page for explicit questions such as How many units by prepper today?, Which printers are offline?, or Which mission reports await approval? It supplies scope, capture time and a verification path; permission is checked again before bridge dispatch. No arbitrary SQL, updates, form filling, approval, printing or rewards. Only technical names/quantities/status are shared; secret fields and private notes are excluded. Open the verification report to review full groups; bot context is bounded to 100 groups. Extension v1.0.9 adds View verified operational report links; upload its source and prebuilt ZIP with the site, then users download/unzip/reload it from Browser Extensions to see this link. Existing extensions still receive server facts; no Windows worker update is needed. For a different time range use Scan Activity. Missing permission/data must be explained, not guessed.

Operational Tasks and prepper alerts implemented locally 2026-10-06: scanner/prepper navigation → Tasks (index.php?view=scanner_missions) lists open Productivity SKUs and independent Supply Runs even with Warehouse Legends OFF or an excluded player. Existing Supply page is still supply requests. A known active remembered prepper is required. Each section shows up to 50 open items, descriptions/needed quantities, optional notes, report controls and recent review status. Can’t find it / I restocked it creates an admin-reviewed report using the existing one-pending-report guard and reviewer bell preferences. If game rewards are eligible, unlock the Legends PIN before reporting; game-OFF/excluded operational reports need no game PIN and are permanently ineligible for rewards, even if Legends is switched on later. Approval while game OFF or currently excluded also grants no game rewards; there is no later automatic backfill. Confirm/Deny retain existing mission-review permissions. Submission never awards anything. No inventory restock adjustment is performed by a supply claim. New schema0101 adds reward eligibility; next schema0102, next seed0071. GET/POST scanner-missions-api.php exposes only task/recent-own-report data, uses no-store, remembered-name identity and POST custom-header/same-origin checks. Operational reviewer notices remain enabled with game OFF; other game notifications still respect the switch. Scanner pages poll open tasks every 60 seconds while visible, independently of game quiet mode. A readable task alert offers View tasks and Dismiss for 12 hours; each item ID is dismissed per prepper on this device in local storage. It reappears after expiry if still open, and new item IDs alert separately. Dismissal never closes the task and Tasks always lists it. Storage-blocked devices display a failure instead of claiming dismissal. Dialogs hide alerts; they resume on a later poll. No sound, background push, SMS or email is added; offline/hidden devices receive updates on the next successful poll. Updated app/knowledge/game/01-warehouse-legends.md and manuals in parity; isolated game-off/no-retroactive-reward and player regression/syntax/knowledge checks pass locally. Physical scanner/hosted acceptance pending.

Where to look & Supply Run pictures (2026-10-07): each Productivity SKU card on Tasks shows a 📍 Where to look box: when the SKU was last prepped, by whom and whether that box is still staged or already sent, plus the last location typed for it (for example Silver Shelf or HAZMAT AREA). "No prep history" means it has never been prepped here. Supply Run cards show a picture when a supervisor has set one (Warehouse Legends → Content Studio → Supply Run catalog → Picture URL, an https:// link); otherwise the 📦 icon.

Achievements & Badges collection implemented locally 2026-10-05: open prepper navigation → Legends → Achievements & Badges. Browse configured badges, achievements and career level tracks with visible descriptions, badge rarity/award requirements, achievement targets, progress, earned status and award time. Search by name/description/code; Show filters All, Earned, Not earned, Badges or Achievements & tracks. Compare with selects an active, non-excluded teammate for side-by-side collection progress; Just me clears comparison. Badges show repeat award counts and latest date. Disabled definitions remain visible if already earned/progressed, labeled Retired; disabled unearned definitions are hidden. Personal daily goals are outside this lifetime collection. Teammate comparison is read-only, requires the remembered active non-excluded player and Game ON, and exposes only display name and whitelisted collection/progress fields. It does not expose PINs, private notes, inventory, hidden words, scoring internals or admin records. GET legends-api.php?action=collection&player= provides this read model; viewing does not require PIN unlock or award anything. Existing Satchel badge summary remains. Definitions are managed by administrators in Content Studio; if empty, enable badge/achievement definitions and verify actual scoring progress. Existing five-item maximum navigation wrapping applies. Curated app/knowledge/game/01-warehouse-legends.md and online/offline manual parity updated. Local read-model, privacy, player regression and syntax checks pass; live browser/mobile acceptance remains pending.

Correction recovery update, verified locally 2026-10-05: schema0098 preserves signed coin and item balances. Record edits/deletions now retain the full revoked amount even when the reward was spent, used or gifted. Negative balances mean outstanding corrected rewards, not spendable coins/items; future matching grants automatically repay them, and restored valid entitlements cancel the deficit once. Game Hub shows Corrected reward balance, and Command Center → Player Progress shows signed coins and Items & Abilities quantities. Gift recipients and already delivered real-world rewards are not retroactively confiscated; recovery is charged to the originating player. Old recovery flags created before schema0098 remain review-only because previously discarded amounts cannot always be reconstructed safely. New balance-backed item notices auto-close once all outstanding balances are repaid; unauthorized actions still cannot spend or gift missing stock. Word Find now replays saved letter inputs after zero-quantity, SKU-change and deleted-record corrections, rebuilding the current word/revealed positions/solved count and adjusting solve points, boss damage and source rewards without random rerolls or duplicate reversals. Legacy key-use history resolves from its recorded source; irrecoverably missing source history stops the event for retry/review instead of guessing. Pallet Pirate now rechecks the original warehouse day/round: losing all qualifying scored units removes daily XP, counter-damage and raid progress; another surviving qualifying record keeps it, and restoring work restores the original amounts once even if settings/assignment later changed. Existing game-enabled/admin permissions and queue retry rules apply. Check Audit Ledger → Points Ledger/Integrity Flags if a correction fails; retry a failed event only after resolving the underlying problem. Browser walkthrough completed with game ON in a disposable database: all 12 Hub tabs, PIN unlock, one-time daily box, scanner trophy chip/HUD, collapse/focus retention and 390px station/Hub layout pass. This does not enable the production game or replace the owner’s operational launch decision.

Player Progress (locally verified 2026-10-05): open Warehouse Legends → Command Center → Player Progress (admin/legends.php#progress). Requires legends.command.view and legends.players.view; the tab and detail API both enforce access. Search by name, rank or status, then View progress. All registered preppers are included, even inactive/excluded or without a game profile. AI Test Prepper is synthetic and does not appear. Profile & Stats shows identity/status, rank, level, XP to next level, coin balance, Clean Run, saved momentum, PIN set/lock status, current-round points/damage, warehouse-day points/XP/scored units, lifetime net ledger totals, undeleted warehouse records/units, current season XP/tier when available, Word Find solved count and Pallet Pirate designation. Stored momentum is historical and may have expired. Today is grounded in the ledger warehouse-day metadata; lifetime sums exclude held entries and retain signed reversals. Warehouse totals use undeleted inventory records and differ from scored units. Seasons follow the existing Game Hub date boundaries. Tracks & Goals, Quests, Badges, Items & Abilities, Roles, Kiosk Claims, Mission Reports and Ledger History show saved rows in separate sections. Progress includes historical period keys and internal metric counters where no named definition exists; definitions with no saved player progress are not listed. Histories paginate 50 rows with Previous/Next; Refresh profile reloads the current section, and roster Refresh reloads the page. Timestamps are UTC. PIN secrets, puzzle answers and private ledger multiplier context are never exposed. This is read-only: no impersonation, awards, resets, PIN changes or profile creation. Use existing Players/Audit tools for permitted changes. Isolated query/held-total/pagination/exclusion/no-write/permission tests and PHP/JavaScript syntax pass; browser/theme acceptance remains pending. No schema or scoring behavior changes.

Warehouse Legends live previews (local implementation, 2026-10-05): use the eye icon beside settings or editor fields, or in content-table actions. The preview shows current saved content or unsaved form values, icons, descriptions, colors, selections and field guidance. For an editable single field, adjust the value inside the preview to update both the preview and the unsaved form immediately. Close with X or Escape; use the original Save button to apply changes. View-only fields remain read-only. Supply Run previews summarize quantity, catalog entry and XP; Power Hour previews show a 100-unit multiplier example before other rules/limits; Pallet Pirate previews summarize designation and daily XP/team counter-damage. These are local illustrations, not an execution of the scoring engine or a promise of awards. Actual results require enabled rules, eligibility and live game state. Existing page/view/manage permissions still apply; no preview API or game writes. File imports and secret accepted mystery answers are excluded. AI generation and live player/ledger simulation are outside preview coverage. JavaScript/PHP syntax checks pass; live browser and all-theme acceptance remain pending.

Verified locally 2026-10-05: open Warehouse Legends → Side Missions → Mission Settings (admin/legends-missions.php#settings). Viewing requires legends.missions.view; saving requires legends.missions.manage. Choose a prepper in Pallet Pirate, set Pallet Pirate XP and damage, then Save side mission settings. None disables the designation. The assignment stays until an admin changes it; Daily Roles and Game Hub display it. With the game enabled and an active round, the selected active, non-excluded prepper earns XP and one pirate_raids progress on their first scored scan each warehouse day. Damage records team counter-attack damage, not damage dealt to the boss. Repeated scans and job retries do not grant another bonus for the same prepper, round and day. Changing the assignment can grant the new holder their own daily bonus. Amounts are saved when earned; later settings edits do not rewrite them. A subsequent correction removes the daily grant if all qualifying work for that day/round is lost; restoration re-earns the original amounts once. If missing, check designation, game/round status, exclusion and whether the scan scored or was held. Isolated scoring/replay tests pass; live browser acceptance remains pending. This change does not enable the game.

Productivity SKU Prep List picker, locally verified 2026-10-05: open Warehouse Legends → Side Missions → Productivity SKUs (admin/legends-missions.php#victims). With legends.missions.manage, choose Load Prep List, search FNSKU/merchant SKU/ASIN/item/list name, select an item, set Bonus points (1–1000), then Assign selected SKU. Only positive remaining quantities are offered; the server rechecks the selection against the shared Prep List snapshot and blocks missing/tampered selections and duplicate open targets. At most 100 search results appear at a time; refine search for more. The source list is shown on the assigned target. Loading uses the existing GRTC Prep List connection/cache; an unavailable list requires Operations → Prep List troubleshooting. Existing linked SKUs use the canonical local identity; new identities resolve through the existing SKU resolver. Assignment does not remove Prep List quantity or award points; the existing verified-record scoring flow remains responsible for claims. View requires legends.missions.view; manual code entry is replaced by the picker.

Supply Run Catalog, locally verified 2026-10-05: open Warehouse Legends → Side Missions → Supply Run Catalog (admin/legends-missions.php#supply_catalog). With legends.missions.manage, Add, Edit, switch On/Off, or Delete unused entries using Code, Name, Description, Icon and Sort order. This catalog is independent of Operations → Supply Catalog. In Supply Runs, Post a supply run chooses an enabled entry from the game catalog, quantity (1–999) and XP reward (0–10,000). Disabled entries cannot create new runs; existing runs and reports keep their catalog references. Referenced entries cannot be deleted: switch them off instead. Schema0097 copies only Operations supplies already referenced by game runs into independent entries, preserving IDs/names/history; later Operations edits do not propagate. Legacy Import/Reset preserves the new game catalog. Game Hub → Missions and Reports to Review now read this catalog; PIN/claim/reviewer reward permissions remain unchanged. The game stays OFF pending owner launch approval.

Wrapped-pallet quantity approval workflow, verified locally 2026-10-05: Game Hub → Missions → Wrapped Pallets → Number of pallets wrapped (whole number 1–999) → optional Notes → Report wrapped pallets. Each report belongs to its prepper and requests approval for that entire count. No pallet ID is required. Game ON, an unlocked Legends PIN and a non-excluded player are required. Admin bell alerts identify the prepper and count; authorized reviewers use Command Center → Approvals or Side Missions → Reports to Review (legends.missions.review) to Confirm or Deny the whole report. Counts and total reward snapshots appear in the review queue, Your reports and Player Progress. Confirm pays per-pallet XP and boss damage configured in Mission Settings multiplied by the reported count: at 1,000 XP/25 damage, 2 pallets pay 2,000 XP/50 damage and 4 pay 4,000 XP/100 damage. Submission grants nothing. Denial pays nothing. Damage applies only to the original still-active, unexpired round; late approval grants XP only. Retry keys prevent duplicate submission/rewards, and changed count or notes with the same key are rejected. Separate reports may have equal counts; admins verify the work and reject duplicate claims. Notes allow 300 characters. Game OFF/exclusion block confirmation. Schema0100 adds pallet_count defaulting older single-pallet reports to 1 without changing their stored reward totals. Next schema0101, seed0071. Curated knowledge updated in app/knowledge/game/01-warehouse-legends.md with manual/offline parity; quantity scaling, invalid counts, retries, denial and late-round checks pass locally. Hosted deployment remains unverified.

AI Generator locally implemented 2026-10-05: open Warehouse Legends → Content Studio → AI Generator tab (admin/legends-studio.php#ai). Requires legends.studio.use, legends.studio.ai and help.live.view. Choose one to four content types, request one to eight drafts, enter a brief (5 characters minimum, 3,000 bytes maximum), then Generate drafts. Uses the existing configured Live Help/Iron Bot bridge; when disconnected, configure the connection before generating. My recent requests displays only your latest 15 requests; waiting/generating requests refresh every eight seconds, or choose Refresh requests. Your typed form is retained during background refresh. Failed jobs show an LH reference; check connection/Error Logs and submit again. Parser errors, missing draft blocks and invalid field checks are displayed. Expand each draft to inspect its fields and choose Review in Content Studio. Approval is separate under Content Studio → Drafts and requires legends.studio.approve; approval adds content switched off unless the reviewer chooses switch on. Generation does not publish, change players or enable the game. Kiosk drafts are in-game only, with no gift-card value. Repeated submission of the same form token in the same session does not queue twice. Callback and draft writes are transactional and retry-safe. No new migrations. PHP/JavaScript syntax, isolated help/draft workflow and private request-history checks pass; live Iron Bot/browser acceptance remains outstanding.

Round Scheduler clue-and-answer generator added, verified locally 2026-10-05: Warehouse Legends → Rounds & Bosses → Rounds → Plan a round (or Edit a planned round) → Generate clue pack fills Mystery title, Accepted answers and five progressive clues. The local warehouse library offers pallet, scanner, label printer, prep table and shipping scale mysteries, based on the old scheduler workflow. Regenerate to choose another pack; existing entered mystery fields require confirmation before replacement. The selected week, boss, HP and season remain unchanged. Review/edit all clues and answers, then Save round separately. Generation does not create, start, publish or modify a round, reveal existing hashed answers, award rewards or contact Iron Bot/external word services. It works without AI connectivity. Existing legends.rounds.manage permission and POST CSRF checks protect the round_clues endpoint. Saved answers retain existing normalization and hashing; when editing, empty answers retain current hashes until explicit new answers are saved. Clues retain normal points-per-clue unlock rules and the five-clue limit. If generation fails, a message appears beside the button and manual entry remains available. Five-packs/five-clues/non-repeated-consecutive-answer and syntax checks pass; live browser/hosting acceptance remains pending. No migration or scoring changes.

Warehouse Legends Icon Library picker added, verified locally 2026-10-05: editable icon/emoji fields throughout the shared content editors, Icon Library glyph editor and Pirate icon setting now offer Choose icon. Open it and search the library by name, code, tags or emoji; select a tile to fill the field, then save the normal form separately. Manual emoji entry remains available. The picker reloads the current library when opened, so imported entries become available without restarting. Content Studio → Icon Library adds/edits named emoji entries; Import / Export supports the icons content type through the existing JSON workflow. Use the existing import dry-run to validate first, then import successfully to make entries available; AI-generated icon drafts require separate approval. This library stores emoji glyphs, not uploaded image files or arbitrary HTML. The read-only icon_choices API requires a signed-in user with a Warehouse Legends page view permission; it exposes only icon_code, glyph, label and tags. Managing/importing the library and saving content retain their existing permissions and CSRF checks. Disabled/read-only fields cannot be changed with the picker. Selection fires live preview updates without publishing content or awarding rewards. Long glyphs exceeding a field limit are unavailable for that field. Empty results, no library entries and connection/sign-in failures show guidance; typing still works. PHP/JavaScript syntax checks pass; live browser/theme acceptance remains pending. No schema changes.

Core Settings prepper checklist display fixed, verified locally 2026-10-05: Warehouse Legends → Rules & Scoring → Core Settings → Who sees the test bar retains checked names after Save core settings and reload. The saved comma-separated prepper IDs are compared as text against numeric option keys. This also fixes other prepper multiselect settings in the shared admin shell. Existing saved selections remain intact; reload the updated page to see them, and do not resave an apparently empty older page. To change recipients, check/uncheck names and save; unchecking all saves an empty list. Viewing/editing still requires the existing settings view/manage permissions, and the test bar only appears while its separate switch is on. Selected inactive preppers remain listed with the inactive label. No schema, game scoring or visibility-rule changes. Actual template regression verifies multiple numeric IDs remain checked, other names remain unchecked, and an empty list checks none. Live hosting deployment remains unverified.

Open the prepper navigation → Labels (index.php?view=labels). Quantity, Divider every N labels, Printer and Your name use the same rounded, themed fields as other prepper stations; narrow screens stack the quantity/divider fields. Enter or scan SKU / FNSKU, choose 1–500 labels, optional divider interval and printer, then Preview Label → Send to Print Queue. Your name is optional. A configured printer and running Print Station are required to print; offline printers leave jobs waiting. Printer status appears below the selector. Queue status uses theme colors and text labels. This styling update does not change label output or printing limits. Local source and PHP validation: 2026-10-05; browser verification across all 16 themes remains outstanding.

The sign-in page includes a Return to prepper station button that opens the public station. Access request fields are stacked vertically, with each label above its input. Account request bell alerts (2026-10-05): each new pending request alerts active admin-role reviewers holding users.view, users.update and notifications.account_request.receive. Click the alert to open Users & permissions → Review request. Settings → Notification preferences → Account access requests controls this alert; individual Deny on its receiver permission also suppresses delivery. Managers cannot receive these approval alerts. The migration adds one alert per already-pending account to eligible reviewers; rerunning the backfill does not duplicate existing alerts, including read alerts. No email is sent. Refresh the admin page after a request; check notification preferences and receiver/page/update permissions if no alert appears. Approval review update, verified locally 2026-10-05: Review request opens one editable approval form with Username, Email, Default order type, Default home page, Role and individual permissions. Edit these before selecting Approve and activate account; profile changes, role, permission overrides and activation save in one transaction. Validation failures retain entered details for correction and keep the account pending. Active status is set only by approval. The requested password is retained; password resets are available after approval with users.password.reset. Permission section tabs, locally verified 2026-10-05: Users & permissions → Review request and Edit user → Individual permissions group the permission catalog into tabs by site section, with a count on each tab. Select a tab to review Inherit role, Allow or Deny choices. Switching tabs preserves all choices; Approve and activate account or Save individual permissions saves selections from every tab, including hidden sections. Keyboard users can move between tabs with Left/Right arrows, Home and End. Tabs wrap on smaller screens. Without JavaScript, the full permission list remains available. Existing approval and permission requirements still apply. Account removal styling, updated 2026-10-05: Users & permissions displays Remove as a compact outlined danger button aligned with Edit. Selecting Remove still opens the username-confirmation dialog and requires users.delete. Access requests (2026-10-05): on admin/login.php choose Request access, enter desired username (2-80 characters), email, Confirm email and desired password (12-72 characters), then Submit access request. Pending accounts cannot sign in. Passwords are hashed. Limit: five valid submissions per client per 15 minutes. In Admin Tools → Users & permissions (users.view), pending accounts show Pending. An administrator with users.update chooses Review request, selects a role and Inherit role / Allow / Deny for each permission, then Approve and activate account. Approval saves role and permissions together. Other roles cannot approve; ordinary edits cannot activate pending accounts. Remove discards a request with users.delete and username confirmation. Email confirmation checks matching input only; no email is sent. If submission fails, check username availability, matching email and password length; wait 15 minutes after repeated attempts.

KTools 4.0 is the warehouse inventory, prep verification, packaging tracking, and operational control platform for GRTC. It unifies physical floor scanning stations with executive management pipelines in real time.

• Floor Preppers: High-speed FNSKU scanning, case pack templates, bag sizes, master carton tracking, and instant conflict detection.
• Management & Admin: 2D staging, problem resolution, active Amazon orders, shipments, attendance approvals, and supply replenishment.
• External Automations: Bi-directional sync with the live GRTC REST API, Slack webhooks for critical floor alerts, and audit trails.

Every piece of inventory prepped on the floor follows a strictly auditable pipeline from physical receipt to final carrier dispatch:

• Items with physical damage or missing components are routed to Damage Out.
• Uncataloged, orphan, or mismatched cartons are routed to Misfits for investigation.
• Boxes requiring admin review before 2D entry are flagged as Problem Items.

KTools 4.0 is fully PWA-certified with an offline-ready service worker (service-worker.js) and standalone window mode.

• Installing on Zebra / Android Scanner Guns: Tap the 📲 Install App button in the workstation header or Chrome menu > Install App. The app opens in dedicated full-screen mode without address bars.
• Installing on iPad / iPhone: Tap the Safari Share button ⎋ > Add to Home Screen ⊞.
• Hardware Scanner Focus: Input fields use auto-select on click/tap, keeping the virtual keyboard hidden on mobile scan guns while hardware lasers pump barcode strings directly into focus.

Added 2026-10-07 (verified locally in headless Chrome; physical-station acceptance pending upload). Every prepper station page (Add Record, Misfits, Damage Out, Pick List, Supply, Days Off, Message, Edit) watches its connection so nothing looks saved when it wasn't.

• Orange bar, "No connection, saving is paused": the station lost its connection. Stop scanning new items. The bar checks again every 5 seconds and can't be dismissed while the connection is down.
• Red bar, "Not saved": you pressed Save (or Confirm & Submit) while offline. Nothing was sent, no success sound plays, and your entries stay on screen. Press Save again once the bar turns green.
• Red bar, "Save not confirmed" or "server had a problem": a background save (Misfits batch, Damage Out, Pick List) failed partway or the server returned an error. It may or may not have saved, so check Search before saving again to avoid a double entry.
• Green bar, "Connection is back": saving works again. If it says your last save did NOT go through, press Save again. Dismiss it with ✕.
• Full red "❌ Not saved" page: the connection dropped in the middle of a save. Tap Back to the form (entries are usually still there), check Search, then save again.
• Saves are never retried automatically, because repeating a save could create duplicate records. Pages may still open offline from the app cache, but no save goes through until the connection is back.

Files: assets/save-guard.js (bar and save blocking) and service-worker.js (the Not saved page; admin pages are not affected). No permissions, settings or schema.

Each offline period and blocked save is also reported to Admin Tools → Device Usage once the station reconnects (2026-10-07).

The primary workstation used by preppers at tables to record prepped units into warehouse inventory:

• FNSKU Hardware Input: Hardware scanner lasers or manual FNSKU/UPC entry instantly query local records and the GRTC API.
• Camera Scanner: Mobile stations without hardware guns can tap the camera button (📸) to scan barcodes with Quagga.js.
• Duplicate Active Box Warning: If the same FNSKU is already in the active Box Inventory queue, KTools alerts the prepper with a warning banner, direct edit link, and duplicate override confirmation.
• Order Type Routing: Choose between FBA: Amazon, FBA: Hazmat, AWD, WFS, eBay, and Pre-Prep.

• Case Pack Mode: Loads saved packaging templates for the SKU (units per box, box weight, length, width, height). The prepper enters the number of boxes, and total quantity is calculated automatically.
• Individual Mode: Used when packing loose, irregular, or non-template quantities. Preppers enter total units and box count independently.
• Template Creation: When saving an uncataloged case pack configuration, preppers can save it as a new reusable template for all future preppers.

KTools 4.0 tracks physical warehouse supply consumption directly from prep form entries:

• Selected Packaging Supply: Dropdown lists active prep-listed materials (bubble bags, poly bags, tape rolls, custom mailers).
• Master Carton Auto-Matching: Box dimensions (L × W × H) entered on the prep form are automatically compared against master case specifications in the supply catalog. Matching master cartons are logged automatically based on the box count.

When items arrive in a manufacturer carton and are placed back into that exact same carton after prep:

• Check the "Back in Box" checkbox on the prep form.
• This prevents KTools from deducting a new master carton from the supply catalog, maintaining 100% accurate physical supply inventory.

• Hazmat Mismatch Detection: Submitting an FBA: Hazmat box when local SKU Info does not reflect hazmat automatically adds the item to the SKU Hazmat Review queue and logs an audit notification.
• Mandatory Expiration Tracking: Food, supplements, and perishables require an expiration date. Once flagged, subsequent prep submissions require a valid future date before saving.
• Gated Product Warnings: Gated items display an alert banner with pulse highlights to prompt preppers to verify brand authorization.

For items prepped using automated bagging equipment:

• Preppers select the physical bag dimension (e.g. 6x9, 8x10, 10x13, 14x20).
• Saved bag sizes are committed to SKU metadata and automatically synchronized to the GRTC backend via GrtcBagSizeSync.

• Pre-Submit Confirmation: Displays a high-contrast modal summarizing SKU, units per box, total units, box count, and prepper name before committing to the database.
• Synthesized Web Audio: KTools features a zero-asset Web Audio API tone synthesizer (ScannerAudio):
  • Crisp double-beep on valid barcode scan (1760Hz & 2349Hz).
  • Warning chime on hazmat or duplicate detection.
  • Affirmative triad chime on successful save.
  • Low buzz on invalid or uncataloged barcodes.

Tracks uncataloged, mismatched, or found inventory that does not match active warehouse purchase orders:

• Mode 1: Single Item Entry: Barcode scan with instant catalog lookup, candidate listings grid, category selector, stepper quantity, shelf/bay location, and camera photo capture fallback.
• Mode 2: ⚡ Rapid Batch Scanner: High-speed continuous barcode scanning with live session counters, duplicate item warnings, and batch summary log.
• Warehouse Misfits Directory: Real-time filterable directory showing open units, locations, and resolution statuses.

Photo size (2026-10-07): a photo taken when adding a misfit is resized to about 1600 px (around 300 KB instead of 2–4 MB) and turned upright automatically. Lists show a small thumbnail; tapping it opens the full photo. Older photos are shrunk the first time their thumbnail is shown.

Prepper Search camera scanning added, verified locally 2026-10-05: open the station Search chip (index.php?view=search). SKU / Barcode / Title Contains accepts a handheld scanner, a typed title or an identifier. Click the camera button beside it, allow camera permission and point the camera at a Code 128, EAN, UPC or Code 39 barcode. A detected barcode closes the camera and immediately searches saved inventory events while retaining the selected Order Type and Status filters. This is read-only and does not save a prep record. Close, Escape or clicking the overlay cancels scanning; the camera stops on leaving the page. Manual typing remains available on mobile. Camera scanning requires HTTPS or a browser-supported localhost origin, camera permission and the existing Quagga library; if unavailable, use a handheld scanner or type the code and click Search. Search matches saved SKU identifiers and titles with the existing 50-result default. Unknown codes show no matching records. This public station search retains existing access rules and adds no admin permission requirement. PHP/JavaScript syntax and shared knowledge checks pass; physical camera/device verification remains pending.

Records inventory damaged during transit, handling, or prepping:

• Barcode Auto-Lookup: Automatically resolves product titles, thumbnails, and identifiers from scanned barcodes.
• Damage Classifications: Log specific reason codes (Crushed Box, Torn Packaging, Broken Seal, Liquid Leak, Defective Unit, Expired).
• Dispositions: Assign warehouse disposition (Vendor Return, Destroy / Discard, Liquidate, Recycled, Salvage).
• Photo Evidence: Upload or snap photo verification directly from phone or workstation camera.

Photo size (2026-10-07): damage photos are resized to about 1600 px when saved and turned upright; the queue and Today lists show small thumbnails, and tapping one opens the full photo.

Guides floor pickers to pull inventory from warehouse shelves for prepping tables:

• Location Sorting: Groups picking items by bay, shelf, or bin number to minimize walking distances.
• Item Checkboxes: Pickers check off items as they place them on prep carts.
• Printable Pick Sheet: Print clean paper pick sheets for bulk pulling.

Show 50 more (2026-10-07): the grid shows the first 50 matching items; tap Show 50 more under the list for the next 50 (the button says how many are not shown). Search, Disposition, Category and Pick Status still search every item, and changing any of them starts again at 50. Picking or putting back an item keeps your place. Item photos load as small thumbnails; tapping one still opens the full photo.

Enables floor workers to request restocking of boxes, tape, poly bags, and packaging gear:

• Hardware Scanner Field: Scan supply UPC barcodes directly to auto-select catalog items.
• Quantity Stepper: Tap − and + buttons for rapid quantity changes.
• Duplicate Request Guard: Warns preppers if an active request already exists for the same item to prevent double-ordering.
• Full-Width Directory: Real-time directory table stacked underneath showing open, ordered, and received requests with priority badges.

Self-service attendance reporting for preppers directly from the workstation:

• Planned Day Off: Request scheduled future time off with start/end date and reason notes.
• Call In: Automatically locked to today's date with read-only timestamping to prevent falsification.
• Conflict Warnings: Informs the worker if team members have already requested the same date while still allowing submission.
• History & Cancellation: Workers can review their submissions and withdraw pending requests before admin review.

• Real-time Proxy: Fetches pending orders and quantities directly from GRTC without exposing external credentials to workers.
• Priority Sorting: High-priority orders float to the top with colored urgency badges.
• Bundle Breakdowns: Multi-pack and bundled listings display nested component parts, quantities, and UPCs.
• Single-Item Printing: Tap 🖨️ on any item card to print a standalone prep sheet without printing the entire list.

Live Help uses restricted Iron Bot agent ktools-help for read-only support. Open Help → Live Help and ask a question. Optional page name/path, SKU and error reference accompany your account and general help snippets; no record lookup or automatic admin actions occur. Each account sees only its own saved conversations. Waiting means the Windows bridge has not claimed the question; answering means it has. Failed questions may be sent again. Start a new conversation after 100 messages; limit 30 questions per hour. Do not include passwords, keys or personal information. The bridge must be running; server-only settings and a shared key enable it. The OpenClaw gateway stays private. History is preserved through legacy reset/import. Setup: docs/LIVE_HELP.md. Connection status also appears next to Send question. Check connection again retries loading. Requests time out after 15 seconds and retain your typed question. If website configuration is off or missing its bridge key, sending stays disabled and the status explains the missing setting. Configuring Iron Bot on its PC does not enable the website. Page context has help tips: hover over or focus the question-mark buttons or fields. Page name is the screen heading (for example On 2D Draft); page path is its admin filename (admin/on-2d-draft.php). A full URL is accepted but reduced to an existing admin filename without query values. SKU and ADM error reference are optional identifiers only: they do not fetch item records or logs. General questions can leave the default Live Help context unchanged. Live Help includes verified order-creation instructions: Orders → Active Orders → New Order; enter Order Name, choose Order Type, optional Notes, then Create Order. The New Order button requires orders.create. Send To 2D can assign completed rows to the new order. Answers are requested as plain text for the chat display. Live Help now uses shared system and game knowledge in app/knowledge plus the full online manual topic catalog. Each question receives relevant current excerpts so replies can explain prerequisites, detailed steps and troubleshooting. In-progress game features are labeled separately from verified behavior. This is static guidance, not live record access; undocumented questions require clarification. Every agent must update shared knowledge and both manuals when features or game rules change. Failed questions display an LH reference; accounts with Error Logs permission can use View error details. Error Logs has a Live Help bridge errors source covering authentication/configuration rejection, unclaimed questions, answer timeouts and reported bridge failures. Earlier failed questions appear as historical failures with the original cause unavailable. Keys, claim tokens and question text are not logged. The bot receives your question, account name/role, recent conversation history and general help snippets along with these clues.

Ask Iron Bot from the bottom dock (2026-10-07): on admin pages, the 🤖 button in the bottom dock (next to Team Chat) opens Iron Bot without leaving the page. It needs help.live.view and uses the same Live Help conversations and Windows worker as Admin → Live Help, so answers wait while the worker is offline (the dot on the button shows online, thinking or a new reply). The panel shows which page you are on; Optional clues lets you add a SKU or an error reference. Opening Iron Bot closes Team Chat and vice versa. It is read-only like Live Help: it explains, it never changes data.

Admin → Mailbox → GRTC Email uses the same mailbox account as the old system. Folders show unread counts. Search sender, subject and message text, filter unread mail and page through results. Choose a message to read it below or beside the list; fullscreen and the reader layout controls help on smaller screens. Opening a message marks it read when you have email-management permission.

Select messages to mark read/unread, star/unstar, move to a folder or move to Trash. New folder creates a subfolder under the selected parent. Permanent Trash deletion is not offered. Incoming attachments download from the reader. Email previews block remote images and active content.

Compose, Reply, Reply all and Forward open a centered message dialog. Enter comma-separated To/CC addresses, a subject and message; review and confirm before Send email. Trailer Pickup fills the same Estes recipient, CC, quote and signature as the old template after you enter PRO number and sender name. Text forwarding does not attach original files. A retry of the same submitted message uses its send token to avoid a second delivery in the current session; if delivery succeeded but the Sent copy failed, do not resend.

Mailbox → Worker Messages lists messages from preppers with search, unread filtering, paging, full-message preview, read/unread controls and selected-message removal. My messages is the default. All administrators requires a separate permission. Removed messages retain a recovery snapshot under Data Audit & Restore. Global search includes worker messages addressed to your own account.

The email page requires PHP IMAP on the hosting server and SMTP access. Connection settings live in protected app/config/mail.php (copied from the old mail_config.php); the old installation is a read-only configuration fallback if that file is absent. Credentials are never displayed in the browser or stored in database backups. Email view, sending, management and viewing other administrators’ worker messages have separate permissions. Reading and changing email acts on the shared mailbox used by both versions.

Admin Tools → Data Optimization retains the original separate Orphan Row Finder and Possible Test Data scans. Select a table or scan all operational tables; set a maximum of 10–500 results. Test terms are comma separated; product title fields are excluded. Review matching fields, row data and deletion mode before cleanup. Results have independent pagination.

Data Relationship Flow displays the current database schema, including tables outside cleanup scope. Click tables to expand columns, search for a table, zoom, drag to pan or use fullscreen. Focus a cleanup row to inspect its relationship values and matching related records. Arrows point toward parents.

Separate Damage Out and Misfit image galleries support previews, refresh, individual cleanup, select all eligible and bulk cleanup. Damage-out images linked to active records are protected. Images from deleted items must pass the selected grace period; misfit cleanup lists only unreferenced files. Removed images go to protected recovery storage.

Data Audit & Restore has table count tiles, Audit Trail and Restore Data tabs. Recoverable Data Audit shows actor, date, action and before/after values, including existing deleted rows without maintenance history. Attendance Workflow Audit shows the latest 80 request events. Restore Data groups records by table with independent pagination. Original IDs are preserved; attendance status is recovered from its deletion audit when available.

Cleanup requires a reason and confirmation, records a recoverable snapshot and blocks stale rows and referenced hard deletions. Restore dependencies first. Hard-deleted records without a snapshot require Backup & Restore. Separate permissions control viewing, cleanup and restoration. These pages act only on the current KTools 4 installation. Scans show at most 500 results; restore lists show up to 500 per table, and audit shows the newest 200 matches.

The central command cockpit for warehouse operations managers:

• Queue Meters: Live KPI counters for Active Boxes, Send to 2D, On 2D Draft, Problem Items, Active Orders, and Shipments.
• Prep Notes Flag: When preppers log shortages or overages, the details indicator turns solid blue (📝) for immediate visual triaging.
• Inline Box Inspection: Expand any row to inspect dimensions, weight, prepper name, packaging supplies, and audit timestamps.

Send To 2D and On 2D Draft product previews corrected locally 2026-10-06: hover or keyboard-focus the SKU to see its image and item name. Valid Amazon ASINs use SKU thumbnails; Walmart/non-ASIN items use their saved Walmart UPC or UPC thumbnail. Saved item descriptions provide a fallback name. Untitled records request a read-only, strict-match GRTC product preview on hover, including equivalent UPC padding, within four seconds; existing linked GRTC IDs remain protected. GET admin/send-to-2d-action.php?action=product_preview&id=<box> requires the corresponding page view permission and a current undeleted box on that page. GRTC switch/authentication remain enforced. Preview does not save metadata, alter quantities/status or award scores. Missing catalog matches, unavailable thumbnail files and connection failures may still show an unavailable image/name. Existing scan, edit and Complete actions are unchanged. Verification: PHP/JS and knowledge checks; hosted hover/mobile acceptance pending.

API note highlighting revised locally 2026-10-06: Send To 2D and On 2D Draft color the SKU text link red when any saved item note has an API source (api, api2 or API-prefixed workflow sources). The note icon uses its original plain styling. Internal-only notes retain the normal SKU link color; mixed API/internal notes use red. Deleting the last API note restores normal color after refresh. Existing note hover previews, note dialog, SKU copy and product preview controls remain. Accessible labels identify API notes, and red uses the theme rose variable. No permission, note content or scoring changes. Source/JavaScript and knowledge checks pass; hosted visual acceptance pending.

Send To 2D / On 2D Draft expanded item description restored locally 2026-10-06: choose Details beside a box to see Item followed by the saved product title (including FNSKUs), with saved record description as fallback. A separate Record description appears only if it differs from the product title. Missing both values displays Item name unavailable explicitly. Previously the details template rendered only inventory record item_description and omitted the product title already supplied by the repository. SKU, location, order/pack information, image hover, copy and action permissions are unchanged. Values remain HTML escaped. No lookup, scoring or database changes. JavaScript/PHP and knowledge checks pass; hosted visual acceptance pending.

Walmart UPC lookup padding correction, verified locally against GRTC 2026-10-06: scanner and admin SKU lookups now retry equivalent leading-zero UPC/EAN/GTIN forms when the original 8–14 digit code returns a successful empty GRTC search. Original code is searched first; retries share its existing timeout budget, preserve the GRTC on/off switch and authentication, and stop on transport failure. Exact identifier matching and linked GRTC SKU protection remain. Error logging does not disable lookup. Barcode lookup alone does not save prep work or grant rewards. Example: 0858050008374 resolves to the catalog UPC 858050008374, Blokees Transformers Classic Figure 07 Arcee. After deploying, retry the SKU lookup to refresh existing placeholders; this change does not bulk repair historical records. If still untitled, check GRTC switch, transport/authentication errors and the catalog identifier; an empty catalog search differs from a failed request. Hosted acceptance remains pending.

Walmart-only items found by UPC (2026-10-08): GRTC's search does not search Walmart UPCs, so an item that exists in GRTC only as a Walmart listing (for example 0687748833994, "2oz. Bottle Plastic Weld") used to show "Item name unavailable" on Send To 2D / On 2D Draft and did not link on the prepper scan page. KTools now keeps a small private index of every GRTC Walmart UPC and uses it when a numeric code finds nothing, then loads that item by its GRTC id. Leading zeros do not matter (UPC-A, EAN-13 and GTIN-14 spellings match). The index refreshes itself when a code is missing and it is more than 15 minutes old; that one lookup takes about 5 seconds longer. An item added to GRTC moments ago is found on the next lookup after a refresh. Nothing to set up.

Iron Bot browser extension implemented locally 2026-10-06: Admin Tools → Browser Extensions → Iron Bot — KTools Page Help → Download ZIP. Requires admin_tools.extensions.view to download. Unzip and use Chrome/Edge Extensions → Developer mode → Load unpacked → select iron-bot-help, then pin its toolbar icon. Sign in to KTools admin with help.live.view and click the icon on a supported krisdkent.net/ktools/ or localhost/127.0.0.1 /ktools/ or /ktools4/ page. The movable help window uses the current signed-in account and existing configured Windows ktools-help worker, with no bridge credentials in the extension. Drag its header, focus the header/use arrow keys, resize its corner, minimize, close or click the toolbar icon to reopen. Reopen after page navigation. Position alone is stored in extension local storage; chats are saved privately in existing Live Help history. New chat starts a conversation; existing 30 questions/conversations per hour, 100-message history and queued/answering timeout limits remain. Polling is every five seconds while the window/page is visible. Review page context lists up to 40 visible eligible field labels and types sent with the page heading/path. Current field values, select options, tables, page HTML, images and cookies are not collected for the prompt. Password/PIN/key/contact/account-permission/payment fields are excluded; do not type secrets into questions. Context labels and AI output remain untrusted. Optional strict ironbot-field JSON blocks become Highlight field and Use suggestion controls; malformed, oversized, unknown or stale field references cannot become executable actions. Highlight briefly outlines and scrolls to the field. Use suggestion requires an explicit click, only accepts supported editable visible fields, validates select choices and native field limits, and dispatches normal input/change events without clicking Save/Submit. Some site fields may react immediately to those events; review suggestions before applying and review the form before saving. No automatic record lookup, awards, file changes or form submission. GET admin/iron-bot-extension.php returns authenticated session/connection/private chat state; POST start/send requires session, help.live.view, enabled Live Help and CSRF, using existing ownership and retry guards. Supported Chrome/Edge desktop only; extension activation uses activeTab/scripting/storage with no persistent all-site access. No external websites, browser settings pages or cross-origin iframe contents. SITE toolbar badge means unsupported location; sign-in/permission errors require KTools admin session/Live Help permission; waiting means the configured Windows worker must answer. Field actions depend on an actual AI response using the optional protocol; generic fallback help may only return text. Source/package/catalog, field restrictions, sanitization, private chat/retry, ordinary help/draft regression and syntax checks pass locally. Hosted Chrome/Edge drag/highlight/fill acceptance with the real Iron Bot worker remains pending. Distribution includes prebuilt iron-bot-help-v1.0.0.zip for hosts without PHP ZipArchive.

Iron Bot extension session/error correction 1.0.1, verified 2026-10-06: the live admin/iron-bot-extension.php endpoint was checked in the existing signed-in Chrome session and returned authenticated ok=true, enabled=true and Windows bridge connected. Thus the observed generic Sign in message did not establish that the account was signed out. Version 1.0.0 incorrectly treated every non-JSON response as a sign-in failure. Version 1.0.1 explicitly includes existing session cookies only in same-origin KTools requests (never reads or stores cookies), requests JSON, and reports real authentication rejection separately from missing endpoint (404), blocked request (403), server errors, non-JSON output, invalid JSON shape, network failures and timeouts. Raw response bodies/tokens are not displayed. Reconnect rechecks session/connection without clearing the typed question or chat. Update through Admin Tools → Browser Extensions → Iron Bot → Download ZIP; replace the existing extracted extension files, Reload Iron Bot at chrome://extensions or edge://extensions, refresh the KTools page and reopen Iron Bot. Local ZIP is iron-bot-help-v1.0.1.zip; deploy the updated source/dist and catalog guide for the hosted download page. No new browser permissions, authentication relaxation, bridge settings, scoring or database changes. Browser verification confirms the live website session and bridge status; updated installed extension behavior still requires reloading the new files on the user device. Source, package, response-classification, private-chat/ordinary-help regression and syntax checks pass. If a non-JSON/blocked response persists after updating, use the specific HTTP status to check server output/hosting security rather than rotating keys or signing out blindly.

Iron Bot extension HTTP 406 transport correction 1.0.2 implemented locally 2026-10-06: user-device extension requests returned non-JSON HTTP 406 while the same endpoint recognized the signed-in Chrome session during direct page navigation. This suggests request-origin/header handling or hosting security filtering, not a missing sign-in; the exact hosting rule has not been inspected. Version 1.0.2 injects a packaged page-transport.js into the MAIN page world using the existing activeTab/scripting permission. Content UI remains isolated. The fixed same-site admin/iron-bot-extension.php endpoint is requested from the page context with ordinary same-origin credentials, no custom Accept header, normal FormData for POST and existing no-store/timeouts. Bounded correlated DOM events support only authenticated GET state/history or POST start/send; caller-supplied URLs, scripts, methods and unrelated fields/actions are rejected/ignored. No external endpoint, cookies API, bridge credentials, additional browser permissions, automatic field changes or server security exceptions are added. Existing session, CSRF, ownership, limits and help.live.view checks remain. Reconnect remains available. Download iron-bot-help-v1.0.2.zip from Admin Tools → Browser Extensions after deploying updated source/dist, or use the local ZIP; replace the extracted files including the new page-transport.js, Reload Iron Bot at Chrome/Edge Extensions, refresh the KTools page and reopen it. Merely closing/reopening the old window does not replace injected scripts. If HTTP 406 persists, ask the host to inspect the matching security-rule log for admin/iron-bot-extension.php; do not disable security filtering globally or rotate bridge keys. Local tests verify page-origin credentials/normal headers, fixed endpoint, FormData key allowlist, correlation, repeat injection, malformed/unsupported action and host rejection, HTTP 406 diagnostics, existing field protections/private chat and ZIP parity. PHP/JavaScript/shared knowledge checks pass. Real installed v1.0.2 request acceptance remains pending device reload; this update supersedes v1.0.1 transport handling.

Floor Siren Quick Operation implemented locally 2026-10-06: on any admin page with the toolbar, open Quick Operations → Floor Siren. Requires toolbar.view plus the existing legends.command.power_hour permission; unauthorized users do not see the operation and cannot use its endpoint. The centered themed modal shows the current announcement and its ending time in the admin time zone, or an explicit inactive state. Enter Message (1–200 characters), choose Heads-up, Celebration or Urgent and Show for 5 minutes, 15 minutes, 1 hour or 8 hours, then Broadcast. Broadcasting replaces the current siren and uses the same shared settings and Change History as Warehouse Legends → Command Center → Power Hour & Siren. Clear current siren removes the current announcement without requiring a new message; Close/Escape does not save. Successful broadcast/clear refreshes modal status. POST admin/quick-op.php?op=floor_siren requires CSRF and the same permission. Connected eligible Legends displays show the siren on their next normal refresh while Game ON; expired announcements do not show. The modal warns if the game is off. Offline/hidden clients and device quiet-mode behavior remain governed by existing Legends display rules. No new sounds, push notifications, awards, schema or permission keys are added. If no banner appears, check Game ON, expiration, device Legends visibility/quiet mode and connectivity. Seven isolated checks verify visibility, empty/current states, escaping, broadcast, invalid inputs, clear and permission denial; PHP/JavaScript and shared knowledge checks pass. Hosted/mobile visual acceptance pending.

Read receipts (2026-10-07): when a prepper taps the siren on a station to open the full message, that counts as read. With permission legends.siren.reads.view (admin, manager), the Floor Siren Quick Operation and Warehouse Legends → Command Center → Floor Siren card show Read by (names and times) and Working but not read yet (active, non-excluded preppers who saved prep work in the last 12 hours). A new broadcast starts a fresh list. Unread sirens keep flashing on stations until opened.

Quick Links and Quick Operations menu layering updated locally 2026-10-06: in the admin top toolbar, choose Select Quick Link or Select Quick Operation. Both searchable dropdowns display above the sticky admin navigation instead of having their first entries hidden behind it. On narrower screens the wrapping toolbar stays in normal document flow while retaining its own positioned layer; menu lists still scroll internally. Existing toolbar.view permission, destination permissions, keyboard controls and operation dialogs are unchanged. If the menu still overlaps incorrectly after deployment, refresh the page to load the updated admin stylesheet. Local source and PHP/knowledge validation completed; hosted/mobile visual acceptance remains pending.

Menu layering fixed again (2026-10-08): the Quick Links and Quick Operations dropdowns open over the white top bar at every screen width; previously everything below the first entry was hidden behind it. If a menu still looks cut off after an upload, hard-refresh the page (Ctrl+F5).

Profile photo and Team Chat avatar display updated locally 2026-10-06: admin Account Settings → Profile picture → Upload a photo → choose JPG/PNG/GIF/WebP → Save profile picture. Existing 10 MB and 25-megapixel upload/server limits and settings.avatar.update permission apply. Photos use fixed square frames with object-fit cover, centered circular cropping and no stretching, including existing portrait/landscape uploads; the source file is unchanged and there is no adjustable crop editor. Keep the subject near the center. Profile preview remains visible on mobile. Team Chat shows the saved avatar beside the sender’s username on each message and in the selected teammate header; preset icons or initials are used when no photo exists, with initials/preset fallback on image-load failure. chat.view still controls visible contacts and conversations; chat.send/chat.read and existing CSRF/private two-party scopes are unchanged. Avatar metadata includes only a validated local image path and fallback text, not arbitrary URLs, file contents or new user permissions. Changes appear on the next normal chat refresh; no schema changes. PHP/JS syntax and knowledge parity verified locally; live hosting/browser acceptance pending.

• Searches Box Inventory, Sent Items, SKU Info, Problem Items, Misfits, Orders, and Shipments simultaneously.
• Instant Restore: Recover soft-deleted boxes or records directly from search results without opening database recovery tools.

Real-time polling notifications system alerting managers to floor events:

• Triggers on new Problem Items, Attendance Call-Ins, Supply Requests, Misfits, and Edit Requests.
• Clicking a notification navigates directly to the affected record and marks the notification as read.

The primary daily operations queue for staging boxes into Amazon 2D barcodes:

• Stacked Filters: Filter by keyword, order type, and prepper simultaneously with dynamic breakdown badges.
• Item Notes Travel: Notes attached to an FNSKU follow the SKU across all administrative workflows.
• One-Click Send: Commits boxes into the 2D draft queue with server-side lockouts.

Boxes in transit to Amazon 2D barcode generation are protected from accidental floor modifications. Preppers must use the Request Edit modal if physical box corrections are required.

Move a box to another order (2026-10-08): on On 2D Draft, open a box's Edit window. Under SKU / Order Type, the Order dropdown shows the box's current order (marked "current") and every active (open) order; orders of a different order type are labelled "different order type". Pick another order and Save Changes to move the box; leaving the current order selected changes nothing. The target order is checked to be still open before anything is saved, and the move is recorded in the box's edit history. Needs permission on_2d_draft.move_order (admin, manager) in addition to on_2d_draft.edit; without it the field is not shown.

Permanent searchable archive of all fulfilled prep boxes, including historical SKU names, carrier shipments, and pallet assignments.

• Isolates prep entries with barcode discrepancies, missing dimensions, or packaging exceptions.
• Problem Resolution: Managers resolve issues by correcting quantities, re-assigning FNSKUs, or returning boxes to staging.

Tracks multi-pallet Amazon FBA and WFS shipments from creation to final carrier bill of lading (BOL) closeout.

Ship Walmart (2026-10-11): packs the Walmart (WFS) rows of an order on On 2D Draft into boxes and pallets, prints a temp label for every box, and gives the numbers Walmart's Shipping step asks for. Admin: Orders → Ship Walmart (under Active Shipments). Preppers: the 🚚 Ship Walmart chip, only for preppers picked in Ship Walmart → Settings.

• Start: Orders tab → ＋ Start build on a Walmart order (or the 📦 Ship Walmart link on On 2D Draft / Active Orders). Rows with no order must be moved to an order on On 2D Draft first.
• Case packs: each case has a checkbox. Checking it creates a closed single-SKU box with the next number (S1, S2…) on the active pallet, weighed with the draft row's box weight. If the draft has no case weight, type it once; it is saved on that draft row.
• Loose units: enter a quantity → Add to box. One SKU in the open box makes an S box, two or more make an M box. Enter the box weight → Close box & add to pallet. A box cannot be closed without a weight. Mixed boxes are labelled M1 (wfs-sku, wfs-sku).
• Boxes table: label, WFS SKU, GTIN and quantity per box, weight, pallet, 🖨 print, ↩ Reopen (loose boxes), 🗑 delete. Numbers are never reused: a deleted or unchecked box keeps its number as void.
• Pallets: pallet weight = box weights + empty pallet weight (default 40 lb in Settings, editable per pallet). Enter height and stackable per pallet; ＋ Add pallet makes the new pallet active.
• Walmart numbers: the summary strip shows Items / Units, Single SKU cases, Mixed SKU cases, Pallets and Total weight; 📋 Copy for Walmart gives one pallet row per pallet. Item details check uses the same columns as Walmart's Item details pop-up.
• Finish: locks the build and keeps it in Box History. It does not change the On 2D Draft rows. Units left, missing pallet heights or no boxes block finishing; an admin with Finish permission can override with a reason. An open box with items always blocks.
• Box History: search by order #, Walmart plan ID, WFS SKU, GTIN, item name, box label (S3, M2) or the scanned temp-label barcode (WM<build>-<label>). Boxes and box contents also appear in the top-bar search.
• Permissions: ship_walmart.view, .pack, .finish, .reopen_finished, .box.delete, .labels.print, .history.view, .settings (admin and manager by default). Station access is the Settings allowlist; preppers can finish only when Settings allows it and never with an override.

Troubleshooting: "No case weight on the draft" → type the case weight. "Only N loose units left" → the rest are case packs or already boxed. "Ship Walmart is not turned on for you" (station) → add the prepper in Settings. A row removed from the draft after packing shows a warning; its boxes are kept.

Calculates units per hour (UPH) for each prepper based on verified prep timestamps:

• Time Windows: View performance by Day, Week, Month, or Year to Date.
• Color-Coded UPH: Green (60+ UPH), Amber (35–59 UPH), Red (<35 UPH).
• Output Trends: Multi-prepper trend lines highlight peak shift output and bottleneck periods.

Granular logs of active prep sessions, bagger runs, and carton pack-out times for labor cost accounting.

Aggregates packaging supply consumption to predict restock reorder points before warehouse runs out of poly bags or boxes.

Compares cumulative monthly prep output against annual warehouse benchmarks (50,000 units/mo goal pace, 1M stretch goal) with comprehensive analytics:

• Top 15 SKUs (Last 12 Months): Horizontal volume ranking with interactive hover cards that follow the cursor in real time, displaying the live Amazon catalog image thumbnail, ASIN, SKU title, total units sent, order count, and active months. Automatically flips position when near viewport edges to prevent clipping.
• Year-over-Year & Month Notes: Historical trend lines with clickable milestone badges and supervisor month annotations.
• Cumulative YTD Pace: Flat average vs. conservative weighted projection modes forecasting end-of-year output.
• Daily Burn Rate: Real-time daily tracking against shift unit targets with color-coded threshold status.
• Shipment Times & Turnaround Analysis: Interactive doughnut chart visualizing shipment duration distribution across 4 turnaround tiers (Express 1–2 days, Standard 3–5 days, Multi-Week 6–10 days, Extended 11+ days) with center average turnaround readout. Features dedicated KPI stat tiles and interactive cards for the Top 5 Fastest and Top 5 Slowest shipments (showing business days open, calendar days, units sent, and estimated labor hours), plus a full drilldown table.

Prepper Reports (reports.php / ?view=reports) provides floor preppers and shift leads with live inventory volume visibility, weekly standings, speed metrics, and cumulative pace tracking:

• Prepper Filtering & Personal KPIs: Select any prepper or auto-detect from station cookie (👤 My Stats) to view personal output (Today, This Week, This Month), units/hour (UPH) speed rate derived from verified prep times, and current weekly leaderboard medal rank.
• Warehouse Volume Summary: Real-time operational KPI tiles detailing Team Output Today, Team Output This Week, Staged Queue volume (waiting for 2D barcode), and boxes currently drafted on active 2D shipments.
• Weekly Leaderboard: Live ranking of all warehouse preppers since Monday with podium awards (👑 #1, 🥈 #2, 🥉 #3), box counts, SKU diversity, and animated progress bars measuring volume percentage relative to the shift leader.
• Top SKUs Prepped: High-volume item rankings with live Amazon ASIN thumbnails, FNSKU identifiers, units prepped, and box breakdown.
• Visual Trend Analytics (Chart.js):
  • Daily Output This Week: Dual-series bar charts comparing full warehouse volume vs. individual prepper contributions across Monday through Sunday.
  • YTD Goal Curve: Smooth cumulative volume progression curve plotted directly against annual benchmark paces (50,000 units/mo [600k/yr] and 1,000,000 unit stretch pace) with real-time pace delta readout.
• Warehouse Live Activity: Real-time streaming audit of the latest 12 prepped cartons saved across the warehouse floor with order type badges (FBA / MFN), pack styles, prepper tags, and relative elapsed timestamps.

Managed via AdminAuth with permissions matrix across roles (Super Admin, Ops Manager, Floor Lead, Read-Only):

• Granular capabilities for deleting records, running backups, approving attendance, and executing API calls.
• Session security with automatic timeout and versioned credential hashing (Bcrypt).

Comprehensive OpenAPI specification and interactive in-browser API sandbox (api-sandbox.php):

• REST endpoints for querying SKU inventory, pulling misfits, syncing bag sizes, and dispatching external notifications.
• Scoped API bearer tokens with rate limiting and automated IP audit logs.

KTools 4.0 features a modern CSS design token engine supporting 16 hand-crafted palettes tailored for warehouse lighting conditions:

• Theme Library: Light (high contrast), Dark (OLED black), Bailey (signature violet/purple), Ocean Blue, Midnight, Nord (arctic blue-grey), Slate, Cyber (neon amber), Mint, Forest (deep pine), Lavender, Rose, Crimson, Sunset, Copper, and Mocha.
• Pre-Paint Theme Activation: Stored locally in localStorage.getItem('kt4_scanner_theme') and injected before first paint on the <html data-theme="..."> element to eliminate flash of unstyled content (FOUC).
• Sync Across Modules: Selected theme automatically applies across public stations (Prepper Scanner, Misfits, Damage Out, Manual, Supply Requests, Days Off) and the Admin Cockpit.

Real-time Slack integrations alert floor supervisors and inventory managers when operational intervention is required:

• #warehouse-supplies: Dispatches alerts whenever an Urgent supply request is submitted or inventory drops below the 48-hour safety buffer.
• #prep-hazmat-alerts: Instant notification when a prepper flags a Hazmat discrepancy (e.g. system shows individual poly-bag but item carries an aerosol or battery warning).
• #warehouse-exceptions: Receives damage-out reports with photo evidence, unhandled misfit items, and missing master carton quantities.
• Configuration: Webhook URLs and channel routing can be managed in Admin > Settings > Integrations.

KTools 4.0 provides detailed real-time telemetry for maintaining warehouse hardware and network connectivity:

• Scanner Input Capture: Logs unreadable barcode scan attempts to identify damaged symbologies or misconfigured Zebra/Honeywell HID hardware.
• Database Query Audit: Unhandled SQL exceptions and migration transaction errors are logged to app/logs/database.log with full call stacks.
• Diagnostic Inspector: Accessible to administrators via Admin > Tools > Logs with filtering by log severity (INFO, WARN, ERROR, CRITICAL).

Ensuring zero data loss during peak warehouse operating shifts:

• Automated Nightly Dumps: Scheduled mysqldump runs every midnight, archiving gzipped snapshots of the ktools4 database with a rolling 30-day retention.
• On-Demand Snapshots: Supervisors can generate a timestamped backup before large bulk inventory imports or schema upgrades via Admin > Tools > Database.
• One-Command Rollback: The migration runner supports strict atomic rollbacks via php bin/migrate.php down to safeguard table integrity.

Added 2026-10-07. Admin Tools → Upload Check (permission admin_tools.upload_check.view, admin) answers "did everything actually get uploaded?".

1. On the PC you upload from, open a terminal in the ktools folder and run php bin/check.php. It checks PHP and JavaScript syntax, runs the standard tests (about 2 minutes; --quick takes seconds) and writes app/build-manifest.json.
2. Run php bin/check.php --changed to list the files changed since the last upload (files marked D were deleted: delete them on the server too).
3. Upload those files plus app/build-manifest.json, then open Admin Tools → Upload Check. Complete means every release file is on the server unchanged. Missing files were not uploaded; Different files are older or were edited on the server. Not in this release lists old code files the release no longer includes.
4. When it says Complete, run php bin/check.php --mark-uploaded so the next --changed list starts from here.

Version history: the code is in Git (history kept in C:/xampp/git/ktools.git, outside the website). Every change is a commit that can be viewed or undone on its own. Secrets, uploaded photos and runtime files are not in Git.

Database version status (2026-10-07) also reads migration tracking without applying changes: readiness requires matching release files and zero unapplied database updates. Applied-file checksum differences are informational and ignore FTP line-ending changes. Review before accepting drift; Unknown is not confirmation of readiness. Existing file status Complete describes file matching only.

KTools notes which kind of device each person uses so supervisors can plan scanners, phones and cameras:

• What is recorded: device type (phone, tablet, desktop), system and browser with versions, model when the browser shares it, screen size, touch, number of cameras (camera names only if the camera was already allowed), barcode support, language, time zone and the last IP address. No location.
• When: preppers once their name is set on the device; admins when signed in. Each person can have several devices. A browser or system update keeps the same entry and refreshes its version.
• Counting: every scan lookup (SKU, misfit, damage out, supply, label, case pack, edit, search) and every save adds 1 to that device. Nothing about scanning or saving changes.
• Troubleshooting details: HTTPS, camera permission state, blocked storage or cookies, memory, CPU, graphics, network type and speed, WebView or Enterprise Browser, page load time, the last IP address and which modern browser features work.
• Problem log: script errors, failed requests, files that did not load, camera start failures (with the reason) and successes (resolution, torch), and hardware-scanner timing (whether it ends with Enter). Scanned codes and typed text are never recorded.
• Where to see it: Admin > Admin Tools > Device Usage (permission devices.view): Preppers, Admins and Problems tabs, warnings per device, problem log and device details under each device.

Connection drops (2026-10-07): stations now report Connection lost (offline under 1 minute, 1–5 minutes or over 5 minutes) and Save did not go through (Save pressed offline, connection dropped mid-save, or a server error). They are sent once the station is back online, counted per device and page, and included in the Problems tab and the dashboard Devices tile. One device or spot showing them often points to weak Wi-Fi there; every device at once points to the internet connection or the server.

Warehouse Legends 3 admin management is available under Warehouse Legends: Command Center, Players, Rounds & Bosses, Daily Roles, Career & Progress, Items & Loot, Kiosk & Rewards, Rules & Scoring, Audit Ledger, Content Studio and Game Manual. In addition, the top navigation bar includes a live Warehouse Legends Snapshot icon (🏆) next to the Device Status icon, displaying real-time game status, active season, preppers scoring today, pending queue counts, active boss HP damage meter, daily Titan unit target progress, and recent round results with MVPs.

Open Rules & Scoring → Core Settings to review the game switch. Keep it OFF while the scoring engine and prepper Game Hub/HUD are being completed. The Game Manual shows settings and definitions; mockups do not prove gameplay is live.

Daily engine code has been tested locally for Titan targets, team streaks, random roles, counter-attacks and shipment penalties. Inventory saves, edits, deletions and problem transitions now use a retry-safe scoring queue. Audit Ledger → Points Ledger shows queued/held/failed work. After fixing the underlying error, an administrator with legends.audit.reverse can click Retry on a failed event; the original work snapshot and rules are preserved and Change History records the retry. Game OFF leaves retried work queued. Daily scan coin caps use the original day’s net ledger total. Older releases preserve today’s coin display. Unrecoverable spent or transferred rewards from deleted records appear as reward_recovery Integrity Flags for supervisor review. Prep minutes are optional: blank or 0 saves normally, and productivity speed includes only positive timed sessions. The legacy API retains untimed history for compatibility. Prep-time reports now retain separate saved sessions for repeated SKUs; schema0094 and a fresh legacy import are needed to recover previously omitted history. Scalar quest, track and metric-reward totals now reconcile record edits/deletions in their original period, including removal and later re-earning of rewards. Career level, badge-count and thousands-of-units counters now recalculate from surviving ledger totals, including downward corrections and re-earned rewards. Attendance now distinguishes acknowledged days off from automatically closed requests; call-ins with no work break streaks. Explicit Freeze use is built in the domain; its Game Hub button is still pending. Late work edits/deletions and held releases now recalculate closed Clean Run days and original-period clean-day quest rewards; ordered streak totals rebuild without double counting. Attendance acknowledgement/removal now queues automatic recalculation of affected closed days. Titan crew/Breaker coins and win metrics now reconcile target invalidation, restoration and contributor ranking. Corrected Titan outcomes also rebuild later cached team-streak displays, preserving inactive days and resetting after worked losses. Titan automatic badges and unspent loot now reconcile win invalidation/restoration using recorded outcomes, without chance rerolls. Weekly Titan boxes now reconcile whole-week wins and three surviving work days, preserving the original loot outcome. Historical season rewards now use the matching season date range, preserving newer-season badges during old corrections. Inventory event badges now preserve other qualifying records, while edited/deleted-record loot reconciles per source. First-daily eligibility now moves to the earliest surviving scored record after edits/deletions, without duplicate items. Hidden targets now fill and replace from fresh cached Prep List entries with remaining quantity, without repeating a SKU in the round or making GRTC calls; stale or exhausted caches leave a smaller pool. Corrected Hidden SKU work now retires invalid claims and restores only the original finder when the original SKU returns; deleted or zero-quantity work cannot keep the discovery reward. Boss victory rewards now reconcile corrected work against the original defeat roster and award period, restoring original outcomes without chance rerolls. Corrected boss damage now updates closed round win/loss status and reconciles the following started round’s loss penalty; restored victories reuse original outcomes. Mystery guessing is built in the domain with opening-time checks, hashed answer validation, retry-safe rewards and escalating wrong-answer costs; consumable clue-power items now have retry-safe domain use, consuming one item and adding clue power without XP or damage. Mind Reader now contributes separate clue power on positive work. Actual saved-work tests verify clue contributions disappear on zero edits/deletions, retain original strength on restoration and remain zero after repeated deletion. Quality Guardian’s drawn partner now has a one-use Momentum correction shield; Clean Run and ability-pause consequences still apply. Actual correction-queue tests confirm first-use Momentum protection, retained Clean Run/pause consequences, one audit row and normal reset on the second correction. Categorized Misfits Send to Prep now queues Misfit Medic damage once for an unambiguously matched worker; marks/imported history do not count as resolutions. Actual partial/full Send to Prep saves now pass reward queue tests, including no repeat for an already removed row. Attribution prefers the login account’s linked prepper identity and falls back to one exact normalized name. API prep batches now collect reward audit IDs and queue only after commit; rollbacks cannot award game damage. The full API dispatcher batch create/stage/commit/replay flow now passes game-reward checks while preserving success/idempotent replay responses. Satchel equip/unequip is now built in the domain: one owned enabled ability at a time, no stock consumption, and stale retries preserve newer choices. Gifting now transfers only duplicates, keeps one sender copy, rejects self/excluded recipients and records retry-safe transfer history. Daily Mystery and Weekly Ability box claims now use enabled weighted drops, fixed warehouse periods and recorded outcomes so retries cannot reroll or grant twice. XP and boss-damage consumables now apply configured effects and consume one copy transactionally, with retry-safe results. Word Find now reveals matching FNSKU letters, pays the planned word-length bonus and records replay-safe replacement; saved FNSKU snapshots preserve original letters and Skeleton Key consumes once to reveal one position. Correction reconstruction and player UI remain pending. Calendar claims now require positive surviving scored work today, use the configured day-of-month reward and pay only once. Calendar edits/deletions now revoke the original reward when qualifying work is removed and restore it once when work returns. Spent rewards require supervisor review through recovery flags. Item-specific correction verification, authenticated Rewards UI and multi-step tracks remain pending. Kiosk purchases now verify unlock level, coins and stock, deliver configured in-game items once, and reserve gift-card budget in FIFO order with the monthly player limit. Supervisor Deliver/Refund actions on Kiosk & Rewards → Claim Queue recheck waiting reservations; refunds restore coins once. Month rollover and reduced budgets reallocate unfulfilled cards. Electronic cards must be sent manually. Player Kiosk forms, cosmetics and notifications remain pending. Legends notification choices are now registered in Settings → Notification preferences. New gift-card claims notify authorized supervisors; active accounts linked to a prepper receive that prepper’s card status, Kiosk unlock, badge and gift notices when enabled. Reward rollback and replay do not leave duplicate success notices. Personal Game Hub links and the remaining supervisor event producers are pending. Tier-box odds and remaining effects are still pending. Hosted HTTP acceptance, undo handling, other action hooks, shield browser/player identity checks and authenticated Game Hub/Satchel forms are pending. Historical rounds without defeat snapshots, other shared awards, player celebrations, Season Pass interactions and recovery debts remain in progress. Full game acceptance checks and the prepper screens are still in progress. Targets use warehouse dates; days with no work preserve the team streak, held work is excluded, and retrying daily work does not repeat penalties or rewards.