agentleFS
mannobeatsmannobeats/hyper-taste/skills/ux-design/SKILL.md

ux-design

Skill1 starsChanged 4 days agoMITView on GitHub

The UX pass for one feature of any app (Mac, web or phone) - what things are called, the flow, where it opens (inline, popover, sheet, panel, window, alert), every state, the words, hover help and accessibility, including features that use AI. Use before building a story that adds a screen, an action or words, when reviewing a built screen's UX, or when the user asks for clearer wording, better names, or whether something should be a modal.

27 other projects use a version of this, 3 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use app-brief, design-from-app and 7 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • StijnusSkill

    app-brief

    One-page brief (docs/BRIEF.md) for a new macOS app, written before any code. Use when the user has an app idea, wants to replace an app they pay for, or the research and clone skills are done.

    Also in 3 projects that use this

  • StijnusSkill

    design-from-app

    Rebuild the look of an app the user likes as a native SwiftUI design kit, by measuring its screenshots (sizes, spacing, type, colors, icons), building its parts, and comparing pixel by pixel until they match. Use when the user names an app whose look they want, shares screenshots of one, or says "make it look like X".

    Also in 3 projects that use this

  • launchapp-devSkill

    layout-grid

    Define responsive layout grid systems with columns, gutters, margins, and breakpoint behavior.

    Also in 3 projects that use this

  • hongyuancSkill

    vertical-slice

    Build and evaluate a production-quality vertical slice before advancing from Pre-Production to Production.

    Also in 5 projects that use this

  • timeplus-ioSkill

    build-and-verify

    Build, compile, run server/client/cluster, execute unit/stateless/SQL tests, verify results, and troubleshoot build/test failures. Make sure to use this skill whenever the user mentions building, compiling, ninja, cmake, running tests, test failures, starting the server, or pre-commit checks, even if they just say 'it doesn't build' or 'the test broke'.

    Also in 3 projects that use this

  • CrystalPlatformsSkill

    agent-cli

    Design and audit CLIs that AI agents (Claude Code, Codex, automated pipelines) consume. Applies ten agent-native principles in two tiers — Tier 1 table stakes (non-interactive default, structured output, errors that enumerate, safe retries, bounded responses) and Tier 2 compounding (cross-CLI vocabulary, three-layer introspection, async-aware execution, persistent profiles, two-way I/O) — with a Blocker / Friction / Optimization severity rubric across six command classes (read, mutate, stream, bootstrap, async, config). Use when building, reviewing, or specifying a CLI for agent invocation, mentions argparse, click, cobra, clap, commander, yargs, oclif, thor, agent-friendly or agent-native CLI design, --json output, exit codes, idempotent commands, --wait flags, agent-context introspection, persistent profiles, --deliver routing, --force bypass, vocabulary consistency, or repos with cli.py, main.go, src/cli.rs, bin/<tool>. Skip general shell scripting and human-only TUIs.

    Also in 3 projects that use this

  • +1 more from Stijnus

The file

Show its rules and text

We split this skill into the rules it gives an agent

22 rules. 0 are common, 18 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 27 edited versions we read keep these.

  • 18

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run it once per story that adds a screen, an action, or words.

    line 11
  2. Very specific:

    Read docs/BRIEF.md, docs/GLOSSARY.md (create it if missing), the story, and its reference screenshots.

    line 18
  3. Distinctive:

    Every thing in the app has exactly one name, and every name lives in the glossary.

    line 23
  4. Distinctive:

    Use the user's words, not the code's: "Task", not "TaskEntity" or "record".

    line 34
  5. Distinctive:

    Use the platform's words for system actions.

    line 35
  6. Distinctive:

    Check: every noun and verb the story puts on screen is in the glossary, and no glossary term has a second name anywhere.

    line 43
  7. Distinctive:

    Write the flow as one line, from every way in to a clear end.

    line 48
  8. Distinctive:

    Every step has a way back: Escape cancels, Command-Z undoes.

    line 57
  9. Distinctive:

    Check: the flow fits on one line, and each step names its way back.

    line 60
  10. Distinctive:

    Ask before closing only when people would lose what they typed.

    line 82
  11. Very specific:

    Check: the UX: block names the surface and the table row that picked it.

    line 89
  12. Distinctive:

    Every view the story touches has words and a look for each state:

    line 93
  13. Distinctive:

    Loading: only when it can take longer than half a second.

    line 96
  14. Distinctive:

    Check: each state has a snapshot page or seed data that shows it.

    line 104
  15. Distinctive:

    Add an ellipsis when the command asks for more input before it acts.

    line 109
  16. Distinctive:

    Follow the platform's case: on the Mac, title case for buttons, menu items and window titles, sentence case for labels and messages.

    line 110
  17. Distinctive:

    Check: read every string aloud.

    line 122
  18. Very specific:

    Every icon-only button gets a tooltip ("Add a task to this list (⌘N)"): what it does, starting with a verb, then its shortcut.

    line 127
  19. Distinctive:

    Check: hover every icon-only control in the running app.

    line 135
  20. Distinctive:

    Every custom control has a label, a value, actions, and feedback.

    line 146
  21. Distinctive:

    Check: you ran the flow keyboard-only and with a screen reader, and the platform's accessibility audit shows no issues for the new views.

    line 151
  22. Distinctive:

    Add the result under the story's acceptance lines:

    line 156

“Common” is measured: kept by at least 6 in 10 of the 27 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

UX design

A feature that works can still confuse people. This pass decides what things are called, how a person moves through the feature, where it opens, and what it says, before any code exists.

Run it once per story that adds a screen, an action, or words. The output is a UX: block in the story in docs/STORIES.md, plus new terms in docs/GLOSSARY.md. To review a built screen, run the same steps against its snapshot and list what fails, most serious first.

0. Read

Read docs/BRIEF.md, docs/GLOSSARY.md (create it if missing), the story, and its reference screenshots.

1. Name it

Every thing in the app has exactly one name, and every name lives in the glossary.

| Term | Means | Not |
| --- | --- | --- |
| Task | One thing to do, with an optional due date | to-do, item, entry |
| Complete | Mark a task done; it leaves the list | check, finish, close |
  • Name the things first (the nouns in the brief), then the actions on them.
  • Use the user's words, not the code's: "Task", not "TaskEntity" or "record".
  • Use the platform's words for system actions. On the Mac: Settings, Quit, Close, Undo, Move to Trash (when it can come back), Delete (when it cannot).
  • One word per idea across buttons, menus, titles, tooltips, and errors.
  • Name a feature for what it does for the person, not the technology. A toggle gets a verb phrase that says what and for whom ("Enhance Dialogue", not "Isolate Vocals"). Say the name in a sentence out loud. Method: references/writing.md, Names.

Check: every noun and verb the story puts on screen is in the glossary, and no glossary term has a second name anywhere.

2. Map the flow

Write the flow as one line, from every way in to a clear end.

Command-N, toolbar +, File > New Task -> title field has focus -> Return saves, row appears -> focus stays for the next task
  • List every entry point: menu command, shortcut, toolbar, context menu, drag and drop.
  • Cut steps: fill in good defaults, ask only for what you cannot guess.
  • Every step has a way back: Escape cancels, Command-Z undoes.
  • The end is visible: the new row, a checkmark, the window closing.

Check: the flow fits on one line, and each step names its way back.

3. Pick the surface

Go down this list and use the first one that fits. Lighter is better.

The featureUse
Changes one value on something already on screenInline (field, toggle, menu in the row)
Shows details of the selection while people keep workingInspector
A few related choices tied to one controlPopover
Input people repeat while they watch the result, or a tool kept open beside the workPanel
One short task that must finish or cancel before going onSheet
A long or self-contained task, or one people keep open beside the main windowNew window
A place people come back to, like a list or a collectionSidebar page
A rare action that destroys data and cannot be undoneConfirmation dialog
A problem people must act on right nowAlert
  • Show one modal at a time. Never open a popover from a popover, or a sheet from a sheet.
  • A modal's title names its task: "Rename Project", not "Edit".
  • Return confirms, Escape cancels, and the confirm button names the action.
  • Ask before closing only when people would lose what they typed.
  • Information that needs no action goes inline, never in an alert.

Details for each surface (SwiftUI code included): references/presentation.md. On the web a sheet is a dialog or a drawer, a panel is a side panel, and the same order holds. On a phone, sheets with detents replace popovers and windows.

Check: the UX: block names the surface and the table row that picked it.

4. Cover the states

Every view the story touches has words and a look for each state:

  • Empty: what it is for, plus the first action as a button.
  • Loading: only when it can take longer than half a second. Show cached content first. A bar when the length is known, with words for what is happening ("Importing 12 photos"), and Cancel when it is safe.
  • Error: inline, next to the problem, with the fix.
  • Too much: long names truncate with the full text in a tooltip; a thousand rows still scroll smoothly.
  • Done: visible feedback in the same frame, and Undo where it applies.

Check: each state has a snapshot page or seed data that shows it.

5. Write the words

  • Buttons and menu items start with a verb: "Add Task", "Export…".
  • Add an ellipsis when the command asks for more input before it acts.
  • Follow the platform's case: on the Mac, title case for buttons, menu items and window titles, sentence case for labels and messages; on the web and Android, sentence case everywhere.
  • Errors say what happened and how to fix it, without blame. Write "Unable to load tasks", never "We're having trouble".
  • Confirmations repeat the action: "Delete 3 Tasks?" with buttons "Delete" and "Cancel", never "Yes" and "No".

Patterns and before and after examples: references/writing.md.

A feature that uses AI (a summary, generated text, a model's suggestion) also follows references/ai-features.md.

Check: read every string aloud. Each one is short, uses glossary terms only, and a new user could say what each button does.

6. Add hover help

  • Every icon-only button gets a tooltip ("Add a task to this list (⌘N)"): what it does, starting with a verb, then its shortcut. Not its name again. Rules: references/writing.md, Tooltips.
  • Hover can reveal secondary actions on a row. The same actions are also in the context menu and on the keyboard.
  • Hover feedback shows at once and never shifts the layout.
  • Information people need is always visible. Hover only adds detail.

Check: hover every icon-only control in the running app. Each one shows a tooltip.

7. Check accessibility

  • The whole flow works with the keyboard alone, in a logical focus order.
  • VoiceOver reads every control with the same words as the screen.
  • Meaning never rests on color alone: add an icon or a word.
  • Text meets the platform's minimum (10 pt on the Mac, 11px on the web) and stays readable with increased contrast and the largest text size.
  • Reduce Motion replaces slides and bounces with fades.
  • Every custom control has a label, a value, actions, and feedback.

The audit steps (SwiftUI modifiers included): references/accessibility.md. On the web: semantic elements first, ARIA only where they fall short, and an automated audit (axe or the browser's).

Check: you ran the flow keyboard-only and with a screen reader, and the platform's accessibility audit shows no issues for the new views.

8. Write the UX block

Add the result under the story's acceptance lines:

UX:
- Flow: Command-N, toolbar +, File > New Task -> title field -> Return saves -> focus stays
- Surface: inline row at the top of the list (changes one value on screen)
- States: empty "No Tasks" with "Add Task"; error "Give the task a title." under the field
- Words: "Add Task", tooltip "Add a task to this list (⌘N)", "Undo Add Task"
- Access: Return saves, Escape cancels; VoiceOver "New task title"

Done when the story has a UX: block, every term it uses is in docs/GLOSSARY.md, and steps 1 to 7 pass on its snapshots.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.