agentleFS
ldaumeldaume/agentic-engineering-harness/skills/design/design-product-ui/SKILL.md

design-product-ui

Skill3 starsChanged 3 months agoMITView on GitHub

Use whenever designing or reworking screens of a SaaS product, app, admin tool, or internal tool - navigation and information architecture, settings, tables and bulk actions, permissions, notifications, search and command palettes, keyboard-first workflows, information density, loading and empty states, undo versus confirm, and AI-native interfaces where people review, correct, reject, or undo AI output (SaaS-Oberfläche, App-Screens, Navigation, Einstellungen, KI-Interface). Not for marketing sites or landing pages (design-web) or dashboards and charts (design-dataviz). Implementation belongs to frontend-craft.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    Put the primary action where the task starts.

    line 21
  2. Distinctive:

    Add keyboard paths for frequent tasks: focus order, shortcuts that do not collide with the browser or screen reader, a command palette when …

    line 30
  3. Distinctive:

    Check accessibility: WCAG 2.2 AA, status not by color alone, focus visible and not hidden (verified) [SRC-A11Y-001] [SRC-PROD-016].

    line 33

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Product Interfaces

Design product screens around the work people repeat, so that frequent tasks are fast for experts and learnable for newcomers, and nothing important is hidden to make a screen look calm. A reduced interface is not better when it hides the information or action the task needs.

Golden path

  1. Name the task, its frequency, and who does it (expert, occasional, first-time), because density and discoverability trade against each other and frequency decides which wins.
  2. Map the information architecture before navigation: what objects exist, how people find them, what they do with them (supported) [SRC-PROD-010].
  3. Put the primary action where the task starts; keep secondary actions reachable, not removed.
  4. Design every state of each view: loading, empty, partial, error, success, no permission, offline. State craft and copy follow frontend-craft.
  5. Choose feedback by duration: response within 0.1 s, progress after 1 s, a progress bar with estimate after 10 s (supported) [SRC-PROD-009] [SRC-PROD-007].
  6. Make destructive actions reversible where possible; confirm only severe ones and name the object (supported) [SRC-PROD-008].
  7. Add keyboard paths for frequent tasks: focus order, shortcuts that do not collide with the browser or screen reader, a command palette when the product has many actions (hypothesis; no primary vendor writeup found).
  8. Check accessibility: WCAG 2.2 AA, status not by color alone, focus visible and not hidden (verified) [SRC-A11Y-001] [SRC-PROD-016].
  9. For AI features, load references/ai-native-interfaces.md.
  10. Validate with task-based tests (design-research) and verify in the browser (design-review).

Load when

  • references/product-golden-path.md - for any product screen: density, navigation, tables and bulk actions, settings, permissions, notifications, progressive disclosure, multi-user context.
  • references/ai-native-interfaces.md - when AI generates, proposes, or applies content or actions.

Rules that change the result

  • Marketing aesthetics (large hero type, sparse layouts, decorative motion) do not transfer to dense product screens, because they trade scan speed for atmosphere on screens people use all day (hypothesis).
  • Progressive disclosure hides advanced options behind a clear control, never options that most people need together (supported) [SRC-PROD-006].
  • Motion in product UI is short and functional; see design-core/references/motion.md.

Done when

Each view has its states designed, the frequent task has a keyboard path, destructive actions are reversible or confirmed by name, AI output is reviewable before it applies, and the accessibility checks pass.

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.