agentleFS
phazurlabsphazurlabs/sumi/skills/screen-flow-patterns/SKILL.md

screen-flow-patterns

Skill58 starsChanged 23 days agoApache-2.0View on GitHub

Taxonomy of 25+ screen types, 15+ user flows, and 25+ UI element deep-dives with layout patterns, component hierarchies, state matrices, and best-in-class references. Use when deciding which screens a product needs and what connects them, before any visual design starts.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before anything is styled, two questions decide the shape of a product: which screens exist, and what connects them.

    line 9
  2. Distinctive:

    A flow is a sequence with an exit at every step: If a step has no way back and no way out, it is a trap, and traps are where support …

    line 18
  3. Distinctive:

    Every screen ships all five.

    line 26
  4. Distinctive:

    Every screen passes through states.

    line 30
  5. Distinctive:

    Offline: No network connection. Show cached data with offline banner.

    line 55
  6. Distinctive:

    Search/Results · Checkout · Onboarding · Login · Sign-up · Forgot Password · Empty State · Error 404 · Error 500 · Maintenance · …

    line 105
  7. Distinctive:

    When documenting or communicating user flows, use these standardized conventions.

    line 126

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Screen and Flow Patterns

Mental model

Before anything is styled, two questions decide the shape of a product: which screens exist, and what connects them. Getting this wrong is expensive in a way visual decisions never are -- you can restyle a screen in an afternoon; you cannot re-architect a flow after it ships without breaking people's habits.

  • Name the screen type before designing it. Nearly every screen is an instance of a known type with known states and a known failure mode. "A page where users manage their team" is a List/Feed with an invite flow attached, and saying so imports thirty solved decisions.
  • A flow is a sequence with an exit at every step. If a step has no way back and no way out, it is a trap, and traps are where support tickets come from.
  • Design the states, not the screen. The happy path is the easy fifth of the work. The other four fifths are empty, loading, error and partial.
  • Density is a decision, not a consequence. Pick a tier and hold it.

Constants

Every screen ships all five. A screen specified only in its ideal state is not specified.

Every screen passes through states. Model them explicitly to ensure all states are designed.

6.1 The Five Core States

[Initial] -> [Loading] -> [Loaded / Populated]
                |               |
                v               v
            [Error]       [Empty]
                |               |
                v               v
           [Retry -> Loading]  [Action -> Loading]
  1. Initial: Screen is mounted but no data request has fired yet. Often invisible (immediate transition to Loading).
  2. Loading: Data is being fetched. Show skeleton screens (preferred), spinner, or shimmer. Never show a blank screen.
  3. Loaded / Populated: Data received and displayed. The "happy" state. Most design time is spent here.
  4. Empty: Data request succeeded but returned zero items. Show empty state illustration + CTA.
  5. Error: Data request failed. Show error message + retry button. Distinguish between network error, server error, and permission error.

6.2 Extended States

  • Refreshing: Pulling to refresh or background sync. Show inline indicator without replacing content.
  • Partial Load: Some data loaded, more available. Infinite scroll loading indicator at bottom.
  • Stale: Data is old. Show "last updated" timestamp. Background refresh in progress.
  • Offline: No network connection. Show cached data with offline banner. Queue actions for sync.
  • Permission Required: User has not granted necessary permission (location, camera, contacts). Show explanation + request button.

6.3 State Design Checklist

For every screen, verify you have designed:

  • Loading state (skeleton or spinner)
  • Populated state (with real data, not lorem ipsum)
  • Empty state (zero items, with helpful CTA)
  • Error state (with clear message and retry)
  • Partial/paginated state (loading more indicator)
  • Offline state (if applicable)
  • Permission request state (if applicable)

Section 7: Information Density Levels

7.1 Three Density Tiers

TierDescriptionTarget Elements per ViewportExample Screen Types
LowBreathing room. One idea per fold. Marketing-oriented.3-5Landing, onboarding, login, confirmation, empty state, error pages
MediumBalanced. Content-rich but not overwhelming. Scrolling expected.6-12Profile, detail, settings, blog/article, checkout, contact
HighData-dense. Power users. Scanning and filtering essential.12-25+Dashboard, analytics, file manager, data table, calendar, comparison

7.2 Density Adjustments by Platform

  • Mobile: Reduce density by 30-40% vs desktop. Stack columns. Collapse secondary info behind taps.
  • Tablet: Intermediate density. Use split views to maintain context without overwhelming.
  • Desktop: Highest density acceptable. Multi-column layouts, sidebars, persistent panels.

7.3 Density Controls for Users

  • Comfortable / Cozy / Compact: Gmail-style density toggle. Adjusts row height and spacing.
  • Card / List / Table: View mode toggle. Cards are lowest density, tables are highest.
  • Show/Hide Columns: Let users configure which columns appear in data tables.
  • Collapse Panels: Resizable and collapsible sidebars for user-controlled density.

Section 8: Platform-Specific Screen Conventions

Index

Screen types (32) -- full specification for each in references/screen-type-catalog.md:

Landing · Home/Dashboard · List/Feed · Detail/Show · Profile · Settings · Search/Results · Checkout · Onboarding · Login · Sign-up · Forgot Password · Empty State · Error 404 · Error 500 · Maintenance · Loading/Splash · Notifications Center · Analytics/Reporting · File Manager · Editor/Canvas · Wizard/Multi-Step · Chat/Messaging · Calendar · Map View · Comparison · About/Team · Contact · Blog/Article · Product Gallery · Booking/Reservation · Confirmation/Success

Flow patterns (22) -- full specification for each in references/user-flow-catalog.md:

Authentication · Onboarding · Purchase/Checkout · Search-Filter-Select · CRUD · Invite/Share · Communication · Notification-to-Action · Upgrade/Upsell · Cancellation/Churn · Help/Support · Data Import/Export · Review/Rating · Onboarding Checklist · Collaborative Editing · Subscription/Renewal · Referral · Multi-Device Handoff · Account Deletion · Error Recovery · Settings Management · Content Creation

UI elements (120+) across navigation, input, feedback, overlay and layout -- references/ui-element-deep-dive.md.

When documenting or communicating user flows, use these standardized conventions.

3.1 Node Types

  • Start / End: Rounded rectangle (pill shape). Green for start, red for end.
  • Screen / Page: Rectangle. Contains screen name. Color-coded by section.
  • Decision: Diamond. Contains yes/no or conditional question.
  • Action: Rectangle with rounded corners. User or system action.
  • Sub-flow: Rectangle with double border. Links to separate flow diagram.
  • Wait / Delay: Hourglass or clock icon. Async step (email verification, processing).

3.2 Path Types

  • Happy Path: Solid green arrows. The primary expected journey.
  • Error Path: Dashed red arrows. What happens when things fail.
  • Alternative Path: Dotted blue arrows. Valid but less common routes.
  • Escape Hatch: Gray arrows. Exit points (cancel, back, close).

3.3 Entry Points

  • Direct URL: Globe icon. User arrives via typed URL or link.
  • Deep Link: Chain icon. From push notification, email, or external app.
  • Navigation: Menu icon. From in-app navigation.
  • System Trigger: Gear icon. Automated redirect (session expire, maintenance).

3.4 Edge Case Annotations

  • Conditional: Annotate arrows with conditions (e.g., "if cart > $50").
  • Frequency: Label paths with expected traffic percentage (e.g., "80% of users").
  • Data Dependencies: Note when a step requires data from a previous step.
  • Timeout: Note time limits (e.g., "OTP expires in 10 min").

Section 4: UI Element Pattern Catalog (120+ Elements)

Reference architecture

FileCoversLines
references/screen-type-catalog.md32 screen types, fully specified1731
references/ui-element-deep-dive.md120+ elements, 26 in depth1259
references/user-flow-catalog.md22 flows step by step1116
references/inspiration-reference-guide.mdwhere to look for precedent228
references/transitions-and-platform-conventions.mdpush, modal, tab, deep link; iOS/Android/web84

What every reference file contains

  1. Purpose -- the one job this screen or flow does
  2. The component hierarchy, named
  3. The state matrix: empty, loading, error, partial, ideal
  4. Accessibility requirements specific to the type
  5. Best-in-class examples, named
  6. The failure mode that type is known for

Routing

For deciding which screens a product needs and the full specification of each: read references/screen-type-catalog.md.

For the sequence connecting them -- entry points, steps, decision points, exits and error branches: read references/user-flow-catalog.md.

For a specific element -- navigation, input, feedback, overlay or layout: read references/ui-element-deep-dive.md.

For how screens move between each other and what each platform expects: read references/transitions-and-platform-conventions.md.

For precedent -- where to look and what to look for: read references/inspiration-reference-guide.md.

Cross-References

  • page-composition-engine -- once the screen type is chosen, this composes it
  • navigation-pattern-encyclopedia -- the navigation connecting the screens
  • performance-states-patterns -- the empty, loading and error states in depth
  • ui-pattern-intelligence -- whether the pattern chosen is still current
  • mobile-ux-design -- the mobile form of each screen type

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.