Appearance & desktop shell
Bring theme, wallpaper, Dock/taskbar/Start, menu-bar items and widgets to a described target state.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- System Settings · Windows Settings · GNOME Settings
- Class GUI-need label
G2· GUI-essential
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R01.01.01.01-T01 Swap two Dock icons without disturbing the rest
Original instruction
The Dock holds the seeded app order in fixtures/dock_seed.txt. Swap the positions of FaceTime and Contacts; every other icon, the separator and the Downloads stack must stay exactly where they are. Do not remove and re-add apps via Finder.
- Platforms
mac· macOS- Applications
- Dock · System Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
drag· Drag & drop- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Read com.apple.dock persistent-apps/persistent-others bundle-id lists; must equal seed with the two entries swapped. Terminal withheld; exec audit (eslogger) flags defaults/dockutil/killall Dock -> 0.
Proposed partial credit
50% FaceTime and Contacts swapped; 50% all other 14 items and the stack unchanged in order
R01.01.01.01-T02 Solid dark background and left-aligned, minimal taskbar on Windows 11
Original instruction
Set the desktop background to a solid color #1E1E1E (no picture), align the taskbar to the left, hide the Task View and Widgets buttons, and set Search to 'Search icon only'. Leave every other taskbar setting as is.
- Platforms
win· Windows- Applications
- Windows Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Registry: Wallpapers BackgroundType=1, Control Panel\Colors Background='30 30 30', Advanced TaskbarAl=0, ShowTaskViewButton=0, TaskbarDa=0, Search SearchboxTaskbarMode=1; snapshot diff shows no other changed values. regedit/PowerShell blocked by AppLocker for the agent user; Sysmon EID1/13 detecting reg.exe or PowerShell registry writes -> 0.
Proposed partial credit
30% solid color exact; 30% taskbar left; 20% Task View+Widgets hidden; 20% search icon only and no collateral changes
R01.01.01.01-T03 Brand accent, highlight and auto appearance on macOS
Original instruction
Marketing wants the demo Mac on brand: appearance Auto, accent color Purple, text highlight color set to Other with RGB 255,214,10 via the color picker, sidebar icon size Small, and 'Show scroll bars' Always. Use System Settings > Appearance only.
- Platforms
mac· macOS- Applications
- System Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgment- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
defaults -g: AppleInterfaceStyleSwitchesAutomatically=1, AppleAccentColor=5, AppleHighlightColor RGB within ±0.01 of (1.0,0.839,0.039), NSTableViewDefaultSizeMode=1, AppleShowScrollBars=Always. Terminal and Script Editor withheld; eslogger exec audit detects defaults/osascript, and writes must originate from System Settings, else 0.
Proposed partial credit
20% Auto; 20% accent; 35% custom highlight within tolerance; 25% sidebar size and scroll bars
R01.01.01.01-T04 Rebuild the Windows 11 Start pins with a Dev folder
Original instruction
Start currently has 18 default pins. Make the first page exactly: Edge, Outlook, Teams, File Explorer, Settings, then a folder named 'Dev' containing VS Code and Windows Terminal, then Notepad. Unpin everything else. Folder name must be set via the folder's rename field.
- Platforms
win· Windows- Applications
- Start menu · Windows Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
drag· Drag & drophover· Hover & contexttype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Export-StartLayout run by grader after sign-out/in; pinnedList order and folder (name+children) must match exactly. PowerShell blocked by AppLocker; Sysmon detects start2.bin writes by any process other than StartMenuExperienceHost, or layout imports -> 0.
Proposed partial credit
25% unwanted pins removed; 30% six pins in order; 35% Dev folder with both apps and name; 10% survives sign-out
R01.01.01.01-T05 Match the Dock to a reference screenshot
Original instruction
Make the Dock look like input/dock_target.png: same app set and order, position on the left, auto-hide on, size matching the reference icon pixel size, magnification on, 'Show suggested and recent apps' off, minimize using Scale effect.
- Platforms
mac· macOS- Applications
- Dock · System Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
drag· Drag & dropvis· Visual judgmentgeom· Precise geometry- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
com.apple.dock: persistent-apps order equals target list; orientation=left; autohide=1; tilesize within ±4 of 52; magnification=1; show-recents=0; mineffect=scale. Terminal withheld; exec audit flags defaults/dockutil.
Proposed partial credit
40% app set and order; 30% orientation/autohide/recents/effect; 30% tile size within tolerance with magnification on
R01.01.01.01-T06 Curate the macOS menu bar and Control Center
Original instruction
Configure the menu bar: Bluetooth and Sound always shown, Wi-Fi shown, Spotlight hidden, battery percentage shown, Focus shown only when active, clock analog off with date 'Never' and seconds shown. Then drag the Now Playing item out of the menu bar if present.
- Platforms
mac· macOS- Applications
- System Settings · Control Center
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & drop- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
com.apple.controlcenter NSStatusItem Visible keys for Bluetooth/Sound/WiFi/NowPlaying, Spotlight MenuItemHidden, BatteryShowPercentage, FocusModes=2, com.apple.menuextra.clock ShowDate=2, ShowSeconds=1, IsAnalog=0. Terminal withheld; eslogger exec audit detects defaults/osascript writes -> 0.
Proposed partial credit
30% always-shown items; 20% Spotlight hidden & Now Playing removed; 20% battery & Focus rule; 30% clock format
R01.01.01.01-T07 Place three configured desktop widgets in a precise column
Original instruction
On macOS Sonoma+, add desktop widgets stacked in the top-left column in this order: Calendar 'Up Next' (medium), Weather for Lisbon (small), Reminders list 'Groceries' (large). Widgets must snap to the grid with no overlap; remove any other desktop widgets.
- Platforms
mac· macOS- Applications
- Notification Center · Widgets
- Interfaces
desk· Desktop GUI- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
drag· Drag & dropgeom· Precise geometryclick· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Parse com.apple.chronod/widget desktop state for kinds, sizes and configured intents (Lisbon, Groceries); grader screenshot finds widget frames: x within 40 px of left edge, y order as specified, no overlap. No CLI places widgets; Terminal withheld and eslogger detects defaults/plutil writes to widget state -> 0.
Proposed partial credit
30% three correct widget kinds and sizes; 30% configuration (Lisbon, Groceries); 30% column order and snapping; 10% no extra widgets
R01.01.01.01-T08 GNOME dash favorites, accent and wallpaper
Original instruction
On Ubuntu GNOME, make the dash favorites exactly Files, Firefox, Terminal, Text Editor in that order (remove all others), switch to dark style with the teal accent color, and set ~/Pictures/walls/fjord.jpg as wallpaper for both light and dark styles.
- Platforms
lin· Linux- Applications
- GNOME Settings · GNOME Shell
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
drag· Drag & dropclick· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
gsettings: org.gnome.shell favorite-apps equals ordered list; interface color-scheme 'prefer-dark', accent-color 'teal'; background picture-uri and picture-uri-dark end with fjord.jpg. The gsettings/dconf CLIs are removed from the VM and Terminal launch is blocked during the run (its launcher stays pinnable); an audit log attributes dconf writes to their D-Bus sender, and writes from a shell or script process score 0.
Proposed partial credit
40% favorites exact order; 30% dark style + teal; 30% wallpaper for both styles
R01.01.01.01-T09 Configure a dense Windows taskbar for a dispatcher
Original instruction
A dispatch desk wants: taskbar buttons never combined, taskbar shown on all displays with buttons on the display where the window is open, 'Show badges' off, 'Show flashing' on, and Copilot hidden. Pin Calculator, Snipping Tool and Notepad in that order after File Explorer.
- Platforms
win· Windows- Applications
- Windows Settings · Taskbar
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & drop- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Advanced registry: TaskbarGlomLevel=2, MMTaskbarEnabled=1, MMTaskbarMode=2, TaskbarBadges=0, TaskbarFlashing=1, ShowCopilotButton=0; Taskband pin order via shell folder enumeration. regedit/PowerShell blocked by AppLocker; Sysmon detects reg.exe/PowerShell edits -> 0.
Proposed partial credit
40% combine/multi-display settings; 20% badges/flashing/Copilot; 40% pin order
R01.01.01.01-T10 Pick the on-brand wallpaper by color and set its fit
Original instruction
~/Pictures/Walls has 12 photos. Set as wallpaper the one whose dominant color is closest (ΔE2000) to brand green #2E5E4E per input/brand_guide.pdf, with fit mode 'Fit to Screen' and the fill/background color set to #2E5E4E. Do this on both the macOS and Windows test machines.
- Platforms
mac· macOSwin· Windows- Applications
- System Settings · Windows Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Wallpaper path equals the precomputed closest image (k-means dominant color); macOS com.apple.wallpaper store placement=fit + color; Windows WallpaperStyle=6, Background='46 94 78'. Shells withheld on both VMs; eslogger/Sysmon exec audit blocks defaults/reg.exe writes.
Proposed partial credit
40% correct image chosen; 30% fit mode on each OS; 30% fill color exact
R01.01.01.01-T11 Different wallpapers per Space with a dynamic one on Desktop 3
Original instruction
Create three Spaces on the built-in display. Desktop 1 wallpaper: ~/Pictures/a_ridge.jpg; Desktop 2: solid color Space Gray from Colors; Desktop 3: the built-in dynamic 'Sonoma Horizon' wallpaper with Automatic appearance. Keep 'Show on all Spaces' off.
- Platforms
mac· macOS- Applications
- System Settings · Mission Control
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollvis· Visual judgment- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
com.apple.spaces lists 3 user spaces; ~/Library/Application Support/com.apple.wallpaper/Store/Index.plist maps each space UUID to the expected provider/file; grader screenshots each Space and compares to references (SSIM ≥0.9). Terminal withheld; eslogger detects Index.plist writes by any process other than WallpaperAgent -> 0.
Proposed partial credit
20% three Spaces exist; 25% Desktop 1 image; 25% Desktop 2 solid color; 30% Desktop 3 dynamic wallpaper
R01.01.01.01-T12 Two-monitor Windows desktop with per-monitor wallpapers and taskbars
Original instruction
With a virtual second display attached, set monitor 1 to wallpapers\harbor.jpg (Fill) and monitor 2 to wallpapers\grid.png (Center on solid #000000). Show the taskbar on all displays with buttons only where the window is open, and set the dark Windows mode but light app mode.
- Platforms
win· Windows- Applications
- Windows Settings · File Explorer
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickhover· Hover & contextvis· Visual judgment- Work product
WP08· App/system state- Conditions
hidden· Hidden/stale state- Proposed difficulty
- 5 / 5
Proposed verifier
IDesktopWallpaper::GetWallpaper per monitor ID and GetPosition; Background color; MMTaskbarEnabled=1, MMTaskbarMode=2; Personalize SystemUsesLightTheme=0, AppsUseLightTheme=1. Screenshot of each monitor SSIM ≥0.9 vs reference. PowerShell/regedit blocked by AppLocker; Sysmon detects reg.exe or scripted IDesktopWallpaper COM calls -> 0.
Proposed partial credit
35% monitor-specific wallpapers; 15% positions/background; 25% taskbar multi-display; 25% mixed dark/light mode
R01.01.01.01-T13 Recreate a colleague's macOS desktop from a screenshot
Original instruction
Make this Mac match input/target_desktop.png and its notes: appearance, accent, wallpaper (solid color, sample it), Dock position/size/autohide/app order, desktop icon size and 'Use Stacks' grouping, menu-bar items, and hot corners named in the notes. Only these settings may change.
- Platforms
mac· macOS- Applications
- System Settings · Dock · Finder
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentdrag· Drag & dropclick· Point & clickgeom· Precise geometry- Work product
WP08· App/system state- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Checklist of 14 keys (AppleInterfaceStyle, AppleAccentColor, wallpaper color ΔE2000 ≤3, dock keys, Finder DesktopViewSettings, controlcenter, wvous-*); full defaults snapshot diff must show no other changes. Exec audit blocks defaults/dockutil.
Proposed partial credit
30% appearance+wallpaper; 30% Dock; 20% desktop/Stacks; 20% menu bar + hot corners (per-key credit); -10 pts per collateral change
R01.01.01.01-T14 Save a reusable Windows theme matching a style sheet
Original instruction
Following input/style.md: dark Windows and app mode, transparency off, custom accent #8E44AD shown on Start and taskbar and on title bars, wallpaper slideshow from Pictures\Brand every 10 minutes shuffled, cursor theme 'Windows Black'. Save it as a theme named 'Studio'.
- Platforms
win· Windows- Applications
- Windows Settings · Control Panel
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
%LOCALAPPDATA%\Microsoft\Windows\Themes\Studio.theme parsed: color, slideshow dir/interval 600000/shuffle, cursors scheme; live registry: EnableTransparency=0, ColorPrevalence=1 (DWM+Personalize), AccentColor matches. regedit/PowerShell blocked by AppLocker; Sysmon detects reg.exe/PowerShell writes -> 0.
Proposed partial credit
25% dark modes + transparency; 25% accent exact with surfaces; 25% slideshow settings; 25% Studio.theme saved with all of it
R01.01.01.01-T15 Organize the desktop with Stacks and a precise icon view
Original instruction
The Desktop has 40 mixed files. Turn on Use Stacks grouped by Kind, set desktop icon size 64, grid spacing at maximum, text size 13, labels on the right, show item info on, sort by Date Added. Do not move or rename any file.
- Platforms
mac· macOS- Applications
- Finder
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & drop- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
com.apple.finder DesktopViewSettings IconViewSettings: iconSize=64, gridSpacing=100, textSize=13, labelOnBottom=0, showItemInfo=1, arrangeBy=dateAdded; FXDesktopStacks grouping=kind; Desktop file manifest hash unchanged. Exec audit blocks defaults.
Proposed partial credit
20% Stacks by Kind; 50% icon view keys (equal share per key); 30% files untouched
R01.01.01.01-T16 Tune the GNOME top bar and hot corner behavior
Original instruction
On Ubuntu 24.04 GNOME: show weekday and seconds in the top-bar clock, show battery percentage, disable the hot corner, make workspaces appear on the primary display only, and set 'Activities' overview app grid to not show favorites twice (keep defaults otherwise).
- Platforms
lin· Linux- Applications
- GNOME Settings · GNOME Tweaks
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
gsettings org.gnome.desktop.interface clock-show-weekday/clock-show-seconds/show-battery-percentage=true, enable-hot-corners=false, org.gnome.mutter workspaces-only-on-primary=true; dconf dump diff shows no other changes. Terminal withheld; auditd exec log detects gsettings/dconf use -> 0.
Proposed partial credit
40% clock options; 20% battery; 20% hot corner off; 20% workspaces primary-only and no collateral
R01.01.01.01-T17 Make the Mac shell easier for a low-vision relative
Original instruction
My grandmother finds the screen cluttered and tiny. Keep only Safari, Mail, Photos, FaceTime and Messages in the Dock (plus Finder and Trash), make the Dock large with magnification off, turn off recent apps, set menu bar to Large size, and stop the wallpaper tinting windows.
- Platforms
mac· macOS- Applications
- System Settings · Dock
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
drag· Drag & dropclick· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
com.apple.dock persistent-apps equals 5 bundle ids; tilesize ≥100; magnification=0; show-recents=0; menu bar size key in com.apple.universalaccess/Display; AppleReduceDesktopTinting=1. Terminal withheld; eslogger exec audit detects defaults/dockutil -> 0.
Proposed partial credit
35% Dock app set; 20% large & no magnification; 15% recents off; 30% menu bar size + tinting (intent mapping)
R01.01.01.01-T18 Custom lock screen with no tips on Windows 11
Original instruction
Set the lock screen to Picture using C:\Fixtures\lock_aurora.jpg, turn off 'Get fun facts, tips, tricks and more on your lock screen', set lock screen status to None, and keep 'Show the lock screen background picture on the sign-in screen' on.
- Platforms
win· Windows- Applications
- Windows Settings
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
ContentDeliveryManager RotatingLockScreenEnabled=0, RotatingLockScreenOverlayEnabled=0, SubscribedContent-338387Enabled=0; lock image hash equals fixture via SystemData cache; lock screen status None; grader locks session and screenshots (SSIM ≥0.85 vs fixture). regedit/PowerShell blocked by AppLocker; Sysmon detects reg.exe/PowerShell writes -> 0.
Proposed partial credit
40% picture applied (screenshot); 30% tips off; 20% status None; 10% sign-in background on
R01.01.01.01-T19 Hot corners with modifiers and a Quick Note corner
Original instruction
Set hot corners: top-left Mission Control, top-right Notification Center, bottom-left Lock Screen with Command held, bottom-right Quick Note with Option held. Then verify by triggering Quick Note once and close it without saving text.
- Platforms
mac· macOS- Applications
- System Settings · Desktop & Dock
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickkey· Shortcuts- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
com.apple.dock wvous-tl-corner=2, wvous-tr-corner=12, wvous-bl-corner=13 modifier=1048576, wvous-br-corner=14 modifier=524288; Notes database unchanged (no new note). Terminal withheld; eslogger exec audit detects defaults -> 0.
Proposed partial credit
80% four corner actions (equal share per corner); 20% both modifiers correct
R01.01.01.01-T20 Reach the reference Dock state via terminal (control)
Original instruction
Control pair of T05: with a terminal and dockutil available, bring the Dock to the exact state in input/dock_target.json (app order, left position, autohide, tile size 52, magnification, recents off, scale effect) and restart the Dock.
- Platforms
mac· macOS- Applications
- Terminal · dockutil
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Same verifier as T05 on com.apple.dock keys; used to measure the GUI-vs-CLI gap on identical end states.
Proposed partial credit
40% app set and order; 30% orientation/autohide/recents/effect; 30% tile size and magnification
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.