mobile-app-design
Skill0 starsChanged 10 days agoMITView on GitHub
Designs and builds React Native and Expo app screens from a brief, an idea, a mockup, a screenshot, a Figma frame or HTML. Load it before writing the first screen file, alongside the Expo skills; they cover the APIs, this covers the design. Infers routes, draws tab bars and headers once in the navigator, keeps colours and sizes in one token module, decides each screen's empty, loading and error states, avoids React Native port mistakes that fail silently, and checks screenshots before done. Use when asked to design, build or make a mobile app or its screens, add screens to an app, build dark mode, or turn a mockup, screenshot, Figma frame or HTML prototype into React Native or Expo code, also when a brief exists. Not for websites, native Swift or Kotlin UI, paywall, onboarding or permission screens (paywall-onboarding), restyling finished screens that look generic (mobile-deslop), choosing a look when no brief exists (mobile-design-brief) or a critique with no edits (mobile-screen-review), each if installed.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- podo/design-agent-skillstheir own version, first seen Nov 8, 2026
- fluxnative-app/fluxnative-claudeword for word, first seen Nov 8, 2026
- jasonfesta/design-agentstheir own version, first seen Oct 15, 2026
- noestradamus/skillstheir own version, first seen Oct 1, 2026
- valewnrt/design-skillstheir own version, first seen Jul 24, 2026
Other versions5 projects use a version of this.
Word for word · 1
- fluxnative-app/fluxnative-claude★ 0100% same
Edited · 0
None found.
Wrote their own version · 4
- podo/design-agent-skills★ 32same name
- valewnrt/design-skills★ 2same name
- jasonfesta/design-agents★ 1same name
- noestradamus/skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- app-sample-contentSkill
- fluxnative-mcpSkill
- mobile-design-briefSkill
- mobile-deslopSkill
- mobile-screen-reviewSkill
- paywall-onboardingSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 1 is distinctive, and 5 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 45
If mobile-design-brief is installed, use it, then come back for Build.
- Very specific: line 67
If mobile-screen-review is installed, use it.
- Very specific: line 74
If mobile-deslop is installed, use it.
- Very specific: line 86
If app-sample-content is installed, use it.
- Very specific: line 91
Paywall, onboarding, permissions: if paywall-onboarding is installed, use it.
- Distinctive: line 102
Never report content as missing from a viewport crop: read the element tree or the source first.
“Common” is measured: kept by at least 6 in 10 of the 4 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.
Files
Mobile app design
sniff → brief → build → review → deslop → sample content → verify → report, starting with a Ran on line and ending with a Not verified line
Each step runs here with a short fallback, or in another skill of this pack if installed. Load a reference when its step starts.
References
- build-screens.md: at Build.
- rn-silent-failures.md: at Build when porting a mockup, HTML or a web design, and again before Verify.
- done-checklist.md: at Verify, before you say done.
- verify-without-device.md: when you write the Ran on and Not verified lines.
- fluxnative-apps.md: only when the FluxNative marker matches (last section).
- fluxnative-mcp-calls.md: only when working over the FluxNative MCP (last section).
Sniff
package.json:expoorreact-native, the Expo SDK version, the router, any UI library.- The token file (
src/theme/,theme.ts, a tokens module), a brief (DESIGN.md), andAGENTS.mdorCLAUDE.mdat the root. - Adopt what exists: one styling system, one token module, one icon family. Never add a UI kit or NativeWind. An existing brief wins over this skill; name any conflict in the report.
- No
expoorreact-nativeinpackage.json: ask before creating a project.
Pick the job
| The request | Steps |
| A new app or screens from an idea | 1 to 6 |
| A mockup, screenshot, Figma frame or HTML to code | 1 (route map and tokens from the images), then 2 to 6 |
| One new screen in an existing app | its job, lead and states, then 2, 3 and 6 with the app's tokens |
| A better look, no new screens | 4 only |
| A critique, no edits | 3 only |
1. Brief
If mobile-design-brief is installed, use it, then come back for Build. Otherwise write DESIGN.md, under 40 lines:
- Product and person: one line each: what it does; who opens it, and when.
- Style direction: one paragraph: the mood; the colour logic (a neutral ground, one accent with one meaning, status colours for status only); how the type feels; how components look (radius, surface, separation); one or two details that belong to this product. Personality goes in colour, type and imagery; layout and navigation stay conventional.
- Screens: each screen's job and its lead (the element seen first). Leads differ between screens.
- Route map: each route marked tab, push, modal or sheet.
- Tokens: hex values, at most 6 type sizes, 3 to 4 radius steps, spacing 4/8/12/16/24/32.
2. Build
Load references/build-screens.md. In short:
- Token module, then navigator layouts with the tab bar and headers, then shared parts, then screens in route-map order.
- Screens hold no colour, size or radius literals.
- Each screen's states are decided before its code: zero items, one item, many, loading, error, reopening a saved record.
- Porting from an image or HTML: check every item in
references/rn-silent-failures.md. - From a reference, take structure and spacing only, never its brand, copy or imagery.
For router, native control and animation APIs, use the Expo skills if installed.
3. Review
If mobile-screen-review is installed, use it; it does not edit. Otherwise, on a full-height shot of
each screen in light and dark, write each problem as screen | what is on screen | fix, and check:
one element leads; one filled primary button per viewport; nothing under a bar; no clipped text;
every enabled control does something.
4. Deslop
If mobile-deslop is installed, use it. Otherwise run the Code commands in
references/done-checklist.md, plus these two, and look at each hit:
rg -n '[\x{1F300}-\x{1FAFF}\x{2600}-\x{27BF}]' app src -g '*.tsx' # emoji used as icons: usually 0
rg -n "textTransform:\s*['\"]uppercase" app src -g '*.tsx' # usually at most 2 per screenA hit is a reason to look; you decide.
5. Sample content
If app-sample-content is installed, use it. Otherwise: names common in the app's city or market,
never the generated default cast (Maya, Priya, Theo, Marcus, Leah, Ines, Hana, Jules), uneven
figures, every total derived from the same records, dates computed from today, and no "Lorem",
"John Doe", "Item 1" or "coming soon".
Paywall, onboarding, permissions: if paywall-onboarding is installed, use it. Otherwise: the
billed price is the largest price; close is at least 44 pt and visible in the first frame; Restore,
Terms and Privacy are present; each permission is asked at the moment of use, after one line
saying what it gives.
6. Verify
Tick references/done-checklist.md, or list each unticked item under Not verified.
- Type check, lint and tests first; a clean build says nothing about how a screen looks.
- A show shot (one viewport) for the person; a review shot (the whole screen, top to bottom) to judge from.
- Never report content as missing from a viewport crop: read the element tree or the source first.
- Name what ran (web export, simulator or emulator, device) and claim only what it proves.
- At most 3 rounds of fix, re-render and re-check.
Report
Ran on: <Expo Web 393×852 DPR 2 | iOS simulator <model> | Android emulator <model> | device <model> | code only>
Routes: <route → tab, push, modal or sheet>
States: <screen: zero, one, loading, error, reopened → shown, or not reached>
Checks: <type check, lint, tests>
Hand-offs: <skills used, or fallback>
Files changed: <paths>
Not verified: <every "cannot" for the source, screens not reached, states not shown>Never
- A status bar, notch spacer, home indicator or second tab bar drawn inside a screen.
- "Coming soon", or a control that does nothing.
- Invented data to make a chart or list look full.
- A phone result claimed from a web render.
If this is a FluxNative app
The root AGENTS.md starts with # Generated app source contract and src/platform/, src/theme/
and src/features/ exist: read references/fluxnative-apps.md before Build. Any other project
follows its own AGENTS.md; change the minimum. Over the FluxNative MCP with no local checkout,
read references/fluxnative-mcp-calls.md before Build.
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
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.

