agentleFS
MattieMattie/mattie_skills/skills/quiet-ux/SKILL.md

quiet-ux

Skill1 starsChanged 9 days agoMITView on GitHub

Design or revise UX and docs so beauty, layout, efficiency, and expressive controls guide the user. Correct the agent's tendency to overcaption, overlabel, and fill space with ornamental microcopy. Apply when the user wants a calmer, more human experience.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    Write prose with $write-humanly if available.

    line 32

≈ “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

Quiet UX

Let the design speak first

If the experience uses filler or ornamental microcopy, you have gone astray. Rework the design so it speaks through beauty, layout, efficiency, and expressive controls. Lead people through a design built for their purpose. Deleting a caption is only enough when the experience already works without it; otherwise, fix what the caption was trying to compensate for.

Preserve beauty and visual accents wherever an expert visual designer would support them. Composition, typography, color, illustration, and motion can give an experience character and make it easier to understand. Quiet UX should feel considered and alive.

Phrases such as "No X, no Y: take your time" spend the user's attention announcing qualities the experience should embody. Give people a pace and path that work instead of telling them the experience is relaxed or simple. If a caption sounds like the developer reminding themselves about a feature or user goal, it belongs in our design notes. Make the interface fulfill that intention.

As an agent, you tend to overcaption, overlabel, and explain every field. Treat that impulse as a reason to look again at the design. A stronger hierarchy, clearer grouping, a better control, or a more direct flow may do the work you were about to give another sentence.

Keep words that do real work

Begin with the quiet experience. Keep essential labels, values, status, and copy that helps someone choose, act, understand consequences, or recover from a problem. Preserve necessary instructions and accessible names; expressive controls still need to be understandable to the people using them, including people using assistive technology.

Remove repeated or floating slogans, text that repeats a heading, narration of obvious sections, congratulations for routine actions, and helper lines that merely occupy space. Keep a line when it serves the person's task or the requested experience. Warmth, wit, story, and a distinctive human voice can belong when they contribute something beyond filling a gap.

For example, a file picker with a visible file list and a clearly labeled "Choose files" control rarely needs "Your files, all in one place. No clutter, no stress." A real constraint such as "PDF, up to 10 MB" helps someone act and should remain visible where it matters.

Offer guidance where it helps

For prototypes and instructional artifacts, add contextual help, examples, or a guided view when they meet a real learning need. Use a toggle only when a richer view is useful, never to hide boilerplate. Essential instructions, consequences, and status belong in the main experience.

Human prose

Write prose with $write-humanly if available. Avoid sounding like a marketing brochure or a technical manual. Sound like a human meant to write this.

Justification: concrete concerns behind quiet guidance

The following user-supplied evidence from September 5-7, 2026 explains concerns about UX created without quiet guidance. Use it to recognize things to minimize: scattered text, design notes inside the product, stacked eyebrow copy and section numbers, repeated slogans, unnecessary helper lines, and template-driven filler. Retain purposeful content and any visual accents an expert visual designer would support; filler still calls for a better design. These reports provide motivation; they do not establish a model-specific cause or whether a particular page was created with quiet guidance.

  • A September 7 Codex subreddit discussion contains the clearest report: Astra's tendency to scatter random text across screens seems stronger. The same thread includes positive reports and recommendations to use reference designs, established components, and design systems.
  • A September 6 Astra interface design thread describes generic, template-driven results. One commenter specifically reports design notes appearing inside the UI.
  • The linked same-prompt gallery provides a concrete example. Its Astra page stacks eyebrow copy, section numbers, repeated slogans, helper lines, and a full FAQ. The test required five sections and invented copy, and its Fable result is also text-heavy, so it does not isolate an Astra-specific cause.
  • A September 6 website-redesign report mentions copy, spacing, color, and layout errors while praising Astra's general brevity.
  • A September 5 frontend thread recommends explicit direction and visual examples instead of vague "make it look good" prompting.
  • A September 5 X discussion mirror argues that frontend design remains an Astra weakness. Replies disagree and recommend having Astra find and emulate appropriate mockups.

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.