Leadup premium ui upgrader
Skill leadupofficial/leadup-claude-skills/leadup-premium-ui-upgrader
Upgrade a LeadUp app's UI to premium international-SaaS standard. Improves layout, spacing, typography, responsive design, dashboard cards, empty/loading/error states, motion, and accessibility while preserving all functionality. Use when the user says "make UI premium", "international standard design", "improve design", "make attractive", "upgrade UI", or "SaaS level UI".From its SKILL.md
npx -y skills add leadupofficial/leadup-claude-skills --skill leadup-premium-ui-upgraderAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
4.9 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it
LeadUp Premium UI Upgrader
Purpose
Raise an app's visual quality to the polish bar of a funded international SaaS — spacing, type, components, the missing states (empty/loading/error), motion, responsiveness, accessibility — without breaking behaviour.
When to use
Trigger phrases: "make UI premium", "international standard design", "improve design", "make it attractive", "upgrade UI", "SaaS level UI", "make it look expensive", "polish the dashboard", "this looks amateur, fix it".
For correctness/console-error testing → leadup-browser-playwright-tester.
This skill is about polish, not functional verification.
Inputs needed
- The app/repo and which screens to upgrade (or "all").
- Any brand constraints (colors, logo, font) or "you choose".
- Framework (Next.js/React, Flutter, Bagisto) for idiomatic changes.
Step-by-step workflow
Use references/premium-ui-checklist.md.
- Audit current UI screen by screen against the checklist; note the worst offenders (cramped spacing, no empty/error states, weak hierarchy).
- Decide a system: spacing scale (4/8px), type scale, restrained palette
- 1 accent, elevation language. Match existing brand if present.
- Prioritize: quick wins (spacing, type, buttons) → components (cards, tables, forms) → the differentiating states (empty/loading/error/success) → motion → responsive → accessibility.
- Implement changes idiomatically for the framework; keep all behaviour, props, routes, and accessibility intact (do not regress a11y).
- Verify responsive at 360/768/1024/1440 and
prefers-reduced-motion. - Document before/after per screen and remaining deeper work.
Required output format
- UI audit — per screen, the issues found.
- Design system applied — spacing/type/color/motion decisions.
- Changes made — per screen, before → after, files touched.
- Prioritized remaining work — quick wins vs deeper.
- A11y statement — confirmed no regression (keyboard/contrast/focus).
Safety rules
See references/security-rules.md. Most relevant here:
- Preserve all functionality, routes, and data flows — visual changes only.
- Do not regress accessibility (keyboard, contrast, focus, labels).
- Do not push/deploy; hand changes back for local review and approval.
Common mistakes
- Restyling that breaks a form submit or route (visual-only means visual-only).
- Adding a flashy theme but still no empty/loading/error states (the real tell).
- Reducing contrast below WCAG AA for "aesthetic" reasons.
- Heavy animations that ignore
prefers-reduced-motionor cause layout shift. - Desktop-only polish on a mobile-first Indian-business SaaS.
- Inventing a brand when one already exists in the codebase.
Troubleshooting
- Under-triggers: user said "this looks cheap" — re-invoke; suggest trigger phrases.
- Over-triggers when they want functional testing → route to
leadup-browser-playwright-tester. - Missing tool/MCP: no preview available — make changes + describe expected result and provide the run command to view locally.
- No internet/browser: proceed; rely on the checklist and code; defer visual confirmation to the user.
- Missing project files: if components can't be found, run
leadup-existing-repo-analyzerfirst. - Build/test failure after changes: revert the offending change, isolate (CSS vs logic), report; never ship UI changes that break the build.
Test prompts
Should trigger (5)
- "Make the jewellery SaaS dashboard premium, international standard."
- "This salon admin UI looks amateur — upgrade it to SaaS level."
- "Improve the design of the clinic site, make it attractive."
- "Upgrade UI for the CRM — proper empty and loading states."
- "Make the school site look expensive and polished."
Should NOT trigger (3)
- "Test that login works in the browser." (→ browser-playwright-tester)
- "Is the payment flow secure?" (→ security-review)
- "Research a UI kit repo on GitHub." (→ github-repo-researcher)
Functional test cases (2)
- Given a dashboard with no empty/error states, the output adds them and lists them under "Changes made" without altering data fetching.
- After changes, the a11y statement confirms keyboard nav + AA contrast preserved and the build still passes.
Success criteria
- Measurable polish against the premium checklist, screen by screen.
- Empty/loading/error/success states present.
- Responsive + reduced-motion respected; a11y not regressed.
- All functionality intact; nothing pushed/deployed.
What ships with it: 2 files
4.6 KB alongside SKILL.md
references/
- premium-ui-checklist.md2.2 KB
- security-rules.md2.4 KB