mcpField
D
design-first-ui-prompting

design-first-ui-prompting

github.com/mengtoDesign & UI7.1k1.8k installsUpdated 3d ago
Repository

About design-first-ui-prompting

A design-first prompting skill for turning vague UI ideas into consistent, detailed specifications. It is for people creating screens with generative models and emphasizes controlled iteration and explicit visual constraints.

Key features

  1. 01Structured UI prompt template
  2. 02One-variable-at-a-time variants
  3. 03Two-pass text workflow
  4. 04Explicit style constraint cards
  5. 05Local visual reference packs

Use cases

  1. 01Drafting a landing page or dashboard prompt
  2. 02Testing color or layout variants consistently
  3. 03Avoiding misspelled text in generated visuals

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/mengto/skills --skill design-first-ui-prompting

After installing, type/design-first-ui-promptingin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

ARTICLE.md
REFERENCES.md
demo/
index.html
PROMPT.md

Details

From the project README

Design-First UI Prompting Skill

This skill is for design-first prompting: turn fuzzy ideas into a tight spec that produces consistent UI.

Core principle

Prompt like a design system, not a wish.

Prompt Structure (copy/paste)

Use this skeleton, then fill the blanks.

GOAL
- What are we making? (e.g., landing page hero / onboarding / dashboard / carousel slide)
- Who is it for? (persona)
- What’s the success criteria? (clarity, conversion, vibe)

FORMAT
- Size/aspect: (e.g., 1080x1350)
- Safe margins: (e.g., 90px)

LAYOUT (wireframe in words)
- Grid: (e.g., Swiss 6-col)
- Placement: (e.g., type-left / image-right)
- Hierarchy: H1 → subhead → body → CTA

TYPE SYSTEM
- Font vibe: (e.g., Söhne / Neue Haas / SF Pro)
- Weights: (H1 700, body 400)
- Leading: (tight for H1, readable for body)
- Tracking: (micro labels wider)

COLOR + MATERIAL
- Background: (hex or description)
- Text: (white/ivory/charcoal)
- One accent only: (cyan/lime/purple)
- Texture: (subtle grain, no plastic HDR)

IMAGERY / UI STYLE
- UI style: (minimal / glass / editorial / playful 3D)
- If photo: lighting + crop + texture rules
- If 3D: materials + lighting + softness

COPY (render EXACTLY)
- Line 1:
- Line 2:
- ...

CONSTRAINTS (change 1–2 things only)
- FONT: ___
- STYLE: ___
- MODE: ___

NEGATIVE PROMPT
- No logos, no watermarks
- No extra text beyond provided lines
- No gibberish typography

Rules that improve consistency

1) Lock one “system”, then iterate with variants
  • First output: nail layout + hierarchy + copy.
  • Variants: change ONE variable at a time:
    • angle / crop
    • accent color
    • card arrangement
    • background tone
2) Treat typography as fragile

If the model keeps misspelling:

  • Use 2-pass workflow:
    1. Generate without text (reserve a clean text-safe area)
    2. Typeset in Figma
3) Use “constraints cards”

When you want the model to obey a style:

  • Add a small “Constraints” panel with explicit values.
  • It anchors the output like a mini style guide.

Example:

Constraints
FONT  CANELA
STYLE  MINIMAL
MODE  DARK
4) Keep a local reference pack

Don’t ask the model to “remember” taste.

  • Save references into a gitignored local reference folder, such as refs/...
  • Point prompts to the reference style

Fast iteration checklist (what to tweak)

  • Spacing: margins, leading, baseline rhythm
  • Contrast: background vs text
  • Hierarchy: one hero line, one support line
  • One accent only (don’t rainbow)
  • Texture: add grain, remove smoothing

Questions to ask (when user is vague)

  • What’s the single message of this screen?
  • What’s the hierarchy (H1 / sub / CTA)?
  • Which style lane: minimal editorial vs playful 3D vs glass UI?
  • Any must-keep constraints (font vibe, color, spacing, grid)?

Comments

Markdown supported