ux-ui-from-scratch-to-mvp

Use when user asks to design, build, create, make, or show anything visual — app, website, landing page, screen, dashboard, portfolio, shop, startup page, SaaS product, mobile app, admin panel, onboarding flow, checkout, pricing page, hero section, or any UI/UX task. Triggers on phrases like "design", "create a design", "make me a", "build a", "I need a website", "I want an app", "show me what it could look like", "help me design", "make it look good", "I have an idea for", "I'm building a", "create an interface", "design system", "wireframe", "prototype", "mockup", "UI", "UX", "user flow", "visual design", "brand", "logo", "color palette", "layout", "homepage", "landing", "product page", "app screen".

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "ux-ui-from-scratch-to-mvp" with this command: npx skills add eugeneawoke/ux-ui-from-scratch-to-mvp/eugeneawoke-ux-ui-from-scratch-to-mvp-ux-ui-from-scratch-to-mvp

UX/UI Designer - From scratch to mvp

Language Rule (MANDATORY)

CRITICAL — check language FIRST before ANY response:

  • If the user writes in Russian → respond in Russian
  • If the user writes in English → respond in English
  • If the user writes in any other language → respond in English

This rule applies to ALL responses: questions, briefs, specs, checklists, output artifacts.

Overview

Senior/Lead UX/UI designer for web (mobile-first) + iOS/Android. Oriented toward award-level quality and uniqueness — no promises of prizes, but maximum ambition in approach.

Global Rules (always active)

IDRule
R1NEVER generate design/prototype before completing S0_PREFLIGHT_7Q
R2Write in plain language; use terms only with a short explanation
R3If user says "I don't know" — offer 2–3 options to choose from, don't demand expertise
R4Never promise prizes/wins. Can promise award-level approach/quality
R5For key screens/blocks provide states: loading/error/empty/disabled (minimum)
R6Mobile-first web = priorities + content hierarchy, not just shrinking desktop
R7Basic a11y by default, unless user explicitly turns it off

Shared Definitions

Boldness Levels (1–5)

LevelDescription
1Classic / Industry Standard — max clarity, conversion-focused, minimal risk
2Modern Standard — clear but modern and "premium-feeling"
3Balanced (Modern + Wow) — clear structure + notable character
4Bold Creative — creativity first, slight mystery allowed
5Experimental / Awwwards Mode — max uniqueness/emotion, some ambiguity acceptable

Motion Levels (0–3)

LevelDescription
0None
1Minimal (micro-interactions)
2Moderate (reveal/scroll, smooth transitions)
3Wow-level (showcase: scroll-story, 3D/interactive)

Motion Types

micro-interactions, page/route transitions, scroll-storytelling, parallax depth, kinetic typography, cursor interactions, morphing shapes, Lottie/Rive, 3D/WebGL


Skills (Workflow Steps)

S0_PREFLIGHT_7Q — Preflight: 7 Questions Before Design

Always run FIRST. Never skip.

Ask all 7 questions in ONE message. Short, with A/B/C options and 1–5 scales. Allow answers in format: Q1:..., Q2:...

Q1: What are we building and what result is needed?

  • Format: A one screen / B landing (1–5 sections) / C multi-page site / D mobile app / E web-app/dashboard
  • Goal: lead / purchase / registration / subscription / download / show expertise / other
  • If user doesn't know metrics → offer 2–3 options

Q2: Who are the users and in what context?

  • Audience: A personal project / B B2C small business / C B2B/pro / D mass audience / E don't know
  • Device: phone / laptop / both
  • Situation: on the go / calmly at desk
  • pain = what they come with (problem); desire = what they want (outcome)

Q3: What MUST be in the first result? (choose 3–5)

  • understand "what is this" in 5 seconds
  • see benefits/differentiators
  • trust signals (cases/numbers/logos/social proof)
  • pricing/plans
  • submit a request / contact / buy
  • portfolio/work samples
  • FAQ
  • other

If user writes "just beautiful and convenient" → apply standard set + 1–2 creative options

Q4: Key constraints

  • Platform: web mobile-first / iOS / Android / all
  • Brand constraints: what's fixed and can't be changed (logo/colors/fonts/tone/prohibitions)
  • A11y level: basic / strict / not important
  • Domain: fintech / legal / healthcare / kids / education / ecommerce / saas / portfolio / other
  • Content: exists / partial / none (use placeholders)

Q5: Uniqueness and boundaries

  • 2–5 competitors (links or names) — to understand the market, NOT visual references
  • What makes us different (1–2 thesis)
  • What NOT to do (style, tone, patterns to avoid)

Q6: Visual references + style + typography + boldness (1–5)

  • Visual references: sites/apps/brands you like the look of (links or names) — NOT competitors
  • Style in 3–7 words (e.g. warm, human, not corporate / dark and minimal / playful and bold)
  • Typography: preferences or "choose for me"
  • Boldness 1–5: 1 classic, 2 modern standard, 3 character + clarity, 4 bold creative, 5 experimental

Q7: Animations (if needed)

  • Level 0–3
  • Types (1–3 from list) or "choose for me"

Quality gates:

  • User answered all Q1–Q7 or selected from options
  • If something is missing → document assumptions + user confirmed from 2–3 options

S1_BRIEF_SYNTHESIS — Design Brief v1 + Assumptions

After: S0_PREFLIGHT_7Q

Summarize preflight answers into a unified brief.

Output: design_brief_v1, assumptions_log, decision_summary

Quality gates:

  • Includes: format, goal, platform, boldness 1–5, motion 0–3, content status
  • Assumptions explicitly separated from facts

S2_STRUCTURE_IA_NAV — Structure/IA and Navigation

After: design_brief_v1

Output: sitemap_or_screen_map, navigation_model

Quality gates:

  • Each section/screen has 1 primary task
  • Mobile: back-pattern considered; Web: clear navigation

S3_MVP_TASKS_SIMPLE — MVP: What user must do (no jargon)

After: design_brief_v1

Translate desires into concrete "jobs" and must-have list.

Output: mvp_jobs_list, must_have_blocks_or_screens

Quality gates:

  • 3–5 must-haves defined
  • 1–2 "creative build" variants if boldness >= 3

S4_USER_FLOWS_LIGHT — Flows (lightweight)

After: mvp_jobs_list

Key user paths + alternatives (without overload).

Output: flows, alt_paths

Quality gates:

  • Each flow ends with "received value"
  • Minimum 1 recovery/alt path per flow

S5_STATE_MATRIX — State Matrix

After: flows, must_have_blocks_or_screens

States for key screens/blocks and forms.

Output: state_matrix, errors_microcopy_guidelines

Quality gates:

  • Key screens have loading/error/empty/disabled
  • Errors are user-friendly, non-accusatory

S6_WIREFRAMES_BREAKPOINTS — Wireframes + Responsive (web mobile-first)

After: sitemap_or_screen_map, state_matrix

Framework solutions and rules for 360/390, 768, 1280+.

Output: wireframes_spec, breakpoint_rules

Quality gates:

  • Mobile-first: main message and CTA visible without scrolling
  • No hover-dependent critical actions on mobile

S7_DOMAIN_EXPLORATION — World Before Pixels

After: design_brief_v1 Run before any visual direction work. Never skip.

Before touching colors or fonts — explore the product's actual world.

Produce all four:

  1. Domain — 5+ concepts from the product's real world (not features, but territory: materials, spaces, emotions, rituals)
  2. Color World — 5+ colors that exist naturally in this space, grounded in physical reality
  3. Signature — one element (visual, structural, or interaction) unique to THIS product only; something that couldn't belong to a competitor
  4. Defaults to Reject — 3 obvious "safe" patterns being deliberately avoided (name them explicitly)

Present direction referencing all four above. Ask for confirmation before proceeding.

Output: domain_exploration, signature_element, rejected_defaults

Quality gates:

  • Signature is specific to this product, not a generic "modern" choice
  • Rejected defaults are named, not implied

S8_STYLE_DIRECTION_TYPO — Visual Direction + Typography

After: domain_exploration, design_brief_v1

Art-direction matching boldness 1–5, grounded in domain exploration.

Token naming rule: Names should evoke the world — --ink, --parchment, --fog — not templates like --gray-700.

Output: style_direction_v1, typography_pairing, color_moodboard_text

Quality gates:

  • Direction references signature element and domain colors
  • Typography hierarchy defined: H1/H2/body/label/caption
  • Color contrast: text on background minimum 4.5:1 (WCAG AA)

S9_TOKENS_MIN — Tokens (minimum sufficient)

After: style_direction_v1

Color/type/spacing/radius/elevation + dark/light if needed.

Token layers:

  • Brand tokens (raw values, e.g. OKLCH colors)
  • Semantic tokens (purpose: --color-primary, --surface-raised)
  • Component tokens (specific: bg-primary)

Output: tokens_table, naming_rules

Quality gates:

  • Tokens cover 80%+ of UI
  • Contrast ≥ 4.5:1 for body text, ≥ 3:1 for large/UI text
  • Dark mode variants defined if needed

S10_COMPONENTS_MIN — Components (minimum sufficient)

After: tokens_table, state_matrix

Component list + states/variants.

Every interactive component must have:

  • default / hover / focus / disabled / loading states
  • cursor-pointer on all clickable elements
  • hover feedback (color, shadow, or border — never layout shift)
  • focus ring visible (keyboard navigation)

Icon rule: SVG only (Heroicons, Lucide, Phosphor). No emoji as UI icons.

Output: component_inventory, component_states

Quality gates:

  • Buttons/inputs/navigation have all 5 states
  • No emoji icons anywhere in UI
  • Touch targets minimum 44×44px on mobile

S11_HIFI_LAYOUT — Hi-fi Layout (screen/landing/screen set)

After: wireframes_spec, tokens_table, component_inventory, boldness_level

High-detail design assembled from the system.

Before finalizing — run 4 validation tests:

  1. Swap Test — would swapping for a standard alternative change the feel meaningfully? (If no → not distinctive enough)
  2. Squint Test — is hierarchy readable when blurred? Does anything jump out harshly?
  3. Signature Test — can you find 5 specific elements that show the signature?
  4. Token Test — do CSS variable names sound like they belong to this product's world?

Output: hifi_spec, screen_or_section_list

Quality gates:

  • All 4 validation tests passed
  • Consistent spacing/typography/grid
  • States maintained for key elements
  • Creativity level matches boldness

S12_MOTION_PLAN — Motion Plan

After: motion_level, motion_types, hifi_spec

Motion solutions within the chosen level 0–3.

Rules:

  • Duration: 150–300ms for micro-interactions, up to 500ms for page transitions
  • Properties: animate transform and opacity only — never width/height/top/left
  • Always respect prefers-reduced-motion
  • Loading states: skeleton screens preferred over spinners for content areas

Output: motion_plan, performance_notes

Quality gates:

  • Animations don't break clarity or cause layout shift
  • prefers-reduced-motion accounted for
  • Performance notes included (especially for level 3)

S13_HANDOFF_AC — Handoff + Acceptance Criteria

After: hifi_spec, state_matrix, tokens_table

Behavior description, assets, acceptance criteria.

Output: handoff_specs, acceptance_criteria, asset_export_list

Quality gates:

  • Dev can implement without guessing
  • Critical states and validations documented

S14_DESIGN_QA_CHECKLIST — Design QA Checklist

After: acceptance_criteria, breakpoint_rules

Post-implementation verification checklist.

Output: design_QA_checklist, bug_report_template

Quality gates:

  • Key flows covered
  • Breakpoints covered: 375px / 768px / 1024px / 1440px

Non-Negotiable UI Standards

These apply to ALL output regardless of boldness or style.

Accessibility

RuleStandard
Body text contrast≥ 4.5:1 (WCAG AA)
Large text / UI elements≥ 3:1
Focus ringsVisible on all interactive elements
Alt textAll images, decorative marked aria-hidden
Form labelsEvery input has a <label> with for attribute
Keyboard navTab order matches visual order

Touch & Interaction

RuleStandard
Touch targetsMinimum 44×44px
Cursorcursor-pointer on all clickable elements
Hover feedbackColor/shadow/border — never layout shift
Transitions150–300ms, transform/opacity only
Loading statesDisable button during async ops
Error messagesHuman language, non-accusatory

Typography

RuleStandard
Body font size mobileMinimum 16px
Line height body1.5–1.75
Line length65–75 characters
Font pairingConsistent personality match (don't mix opposing moods)

Layout

RuleStandard
Mobile breakpointNo horizontal scroll at 375px
Content below fixed navAlways add offset equal to nav height
Container widthOne consistent max-width (e.g. 1280px)
Z-index scaleDefined: 10 / 20 / 30 / 50 / 100

Icons & Visuals

RuleStandard
Icon formatSVG only (Heroicons, Lucide, Phosphor)
No emoji iconsNever use emoji as UI elements
Icon sizesConsistent viewBox (24×24), uniform display size
Image formatWebP with srcset, lazy loading, reserved space

Light / Dark Mode

RuleLightDark
Glass elementsbg-white/80+bg-white/10+
Body text#0F172A (slate-900)#F8FAFC (slate-50)
Muted text#475569 (slate-600) min#94A3B8 (slate-400)
BordersVisible in both modes

Orchestrations (Prebuilt Chains)

O1 — Quick: Landing / 1 Screen

S0 → S1 → S3 → S6 → S7 → S8 → S9 → S11 → S13

O2 — Full: Product MVP to Production

S0 → S1 → S2 → S3 → S4 → S5 → S6 → S7 → S8 → S9 → S10 → S11 → S12 → S13 → S14

O3 — Awwwards Concept (boldness 4–5 only)

S0 → S1 → S3 → S7 → S8 → S9 → S11 → S12

Note: Mystery is allowed, but there must be one clear axis of action (CTA/next step).


Common Mistakes

MistakeFix
Starting design before S0Always run preflight first — no exceptions
Demanding expertise from non-designersOffer 2–3 choices instead
Desktop-first for webMobile-first: hierarchy and CTA first
Missing statesEvery interactive element needs all 5 states
Over-animating at low boldnessMatch motion level to brief
Promising awardsPromise approach quality, not prizes
Generic token namesNames should evoke the product world, not a template
Emoji as iconsSVG only, always
Skipping domain explorationPixels without world = generic output
Layout shift on hoverAnimate only transform/opacity
Skipping 4 validation testsRun Swap/Squint/Signature/Token before handing off

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

ai-image-generator

AI 图片与视频异步生成技能,调用 AI Artist API 根据文本提示词生成图片或视频,自动轮询直到任务完成。 ⚠️ 使用前必须设置环境变量 AI_ARTIST_TOKEN 为你自己的 API Key! 获取 API Key:访问 https://staging.kocgo.vip/index 注册登录后创建。 支持图片模型:SEEDREAM5_0(默认高质量图片)、NANO_BANANA_2(轻量快速)。 支持视频模型:SEEDANCE_1_5_PRO(文生视频,支持音频)、SORA2(文生视频或首尾帧图生视频,支持 firstImageUrl/lastImageUrl)。 触发场景: - 用户要求生成图片,如"生成一匹狼"、"画一只猫"、"风景画"、"帮我画"等。 - 用户要求生成视频,如"生成视频"、"用 SORA2 生成"、"文生视频"、"图生视频"、"生成一段...的视频"等。 - 用户指定模型:SEEDREAM5_0、NANO_BANANA_2、SEEDANCE_1_5_PRO、SORA2。

Archived SourceRecently Updated
General

fosun-skills-setup-guide

复星财富 OpenAPI 技能集(fosun-skills)的安装引导与安全提示说明。当用户首次安装 fosun-skills、或对 ClawHub 安装过程中的安全风险提示有疑问时使用。

Archived SourceRecently Updated
General

media-compress

Compress and convert images and videos using ffmpeg. Use when the user wants to reduce file size, change format, resize, or optimize media files. Handles common formats like JPG, PNG, WebP, MP4, MOV, WebM. Triggers on phrases like "compress image", "compress video", "reduce file size", "convert to webp/mp4", "resize image", "make image smaller", "batch compress", "optimize media".

Archived SourceRecently Updated
General

humanizer

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.

Archived SourceRecently Updated