Design Review logo

Design Review

Community
julianoczkowski
design-review

Run a structured design critique against the brief and codebase. Checks visual hierarchy, consistency, responsiveness, accessibility, and aesthetic fidelity. Use when user wants a design review, critique, QA pass, polish pass, or mentions "review" after building.

Overview

Publisherjulianoczkowski
Repositorydesigner-skills
Skill namedesign-review
Stars
572
Forks
50
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by julianoczkowski on GitHub. Read the source before you install it.

Installation

Install the Design Review AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/julianoczkowski/designer-skills/tree/main/design-review
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/julianoczkowski/designer-skills.git /tmp/designer-skills
mkdir -p .claude/skills
cp -r /tmp/designer-skills/design-review .claude/skills/design-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Design Review in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Design Review on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Design Review is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

This skill runs a structured design review of what has been built, measured against the design brief and the chosen aesthetic philosophy.

CRITICAL — Visual Screenshot Capture

You MUST capture screenshots of the running application as part of every design review. Code review alone is insufficient — you need to see what the user sees. Follow the screenshot capture protocol in Step 3 below. This is not optional.

Example prompts

  • "Review what I just built"
  • "Run a design critique on the landing page"
  • "Check this against the brief"
  • "Here's a screenshot. How does it look?" [paste screenshot]
  • "QA pass before I ship this"

Process

  1. Read the brief. Look for the active feature's brief at .design/<feature-slug>/DESIGN_BRIEF.md. If multiple feature folders exist under .design/, ask the user which feature to review. If no .design/ folder exists, fall back to DESIGN_BRIEF.md in the project root. If neither exists, ask the user what the intended design direction was.

  2. Explore the built code. Examine every component, page, and style file that was created or modified. Scan specifically for:

    • All new or modified components and their relationship to pre-existing components
    • Token/variable usage: are components using shared tokens or hardcoding values?
    • Duplicate components that should be consolidated
    • File naming and organization: do new files follow the project's conventions?
    • Understand what was actually built, not what was planned.
  3. Capture screenshots of the running application.

    This step is mandatory. Do not skip it. Do not rely only on user-provided screenshots.

    Screenshot Tool Priority

    Try each option in order. Use the first one that is available:

    1. Playwright MCP (preferred). Check if the plugin-playwright-playwright MCP server is available. If it is, use it — it gives you precise control over viewport sizing, full-page captures, and file naming.
    2. Cursor IDE Browser (second choice). If Playwright MCP is not available, use the cursor-ide-browser MCP server's browser_take_screenshot tool instead. It has the same core capabilities.
    3. Ask the user (last resort). If neither MCP server nor in-app browser is available, you MUST ask the user to provide screenshots manually. Be specific about what you need:
      • "I don't have access to a browser tool. To complete the visual review I need screenshots of the running application. Please provide:"
      • A full-page screenshot at desktop width (1280px)
      • A full-page screenshot at tablet width (768px)
      • A full-page screenshot at mobile width (375px)
      • Dark mode variants (if applicable)
      • Any specific component or interactive state you want reviewed
      • Ask the user to paste/attach the images directly in chat, or to save them into the screenshots/ folder themselves.
      • Do not skip the visual review. Wait for the user to provide screenshots before proceeding with the checklist.

    Screenshot Save Location

    All screenshots MUST be saved to a screenshots/ subfolder inside the feature's .design/ directory — the same folder where DESIGN_BRIEF.md and other design flow files live.

    Path pattern: .design/<feature-slug>/screenshots/

    If the brief lives at .design/onboarding-flow/DESIGN_BRIEF.md, screenshots go to .design/onboarding-flow/screenshots/. Create the folder if it does not exist.

    If no .design/ folder exists (legacy project or standalone review), fall back to a screenshots/ folder in the project root.

    Use descriptive filenames that encode what was captured:

    .design/
    └── onboarding-flow/
        ├── DESIGN_BRIEF.md
        ├── DESIGN_REVIEW.md
        └── screenshots/
            ├── review-homepage-desktop-1280.png
            ├── review-homepage-tablet-768.png
            ├── review-homepage-mobile-375.png
            ├── review-homepage-dark-mode-desktop-1280.png
            └── review-card-component-hover.png

    Screenshot Capture Protocol

    a. Navigate to the application. Ask the user for the URL if not obvious from the project (e.g., http://localhost:3000). Use browser_navigate to open it.

    b. Capture responsive breakpoints. At minimum, capture these three viewports for every key page/view:

    BreakpointWidth × HeightFilename suffix
    Mobile375 × 812-mobile-375
    Tablet768 × 1024-tablet-768
    Desktop1280 × 800-desktop-1280

    Use browser_resize to set the viewport before each screenshot. Use browser_take_screenshot with fullPage: true to capture the entire scrollable page, and save with the filename parameter pointing to the screenshots/ folder.

    Playwright MCP example sequence (assuming feature slug is onboarding-flow):

    1. browser_navigate → { url: "http://localhost:3000" }
    2. browser_resize   → { width: 1280, height: 800 }
    3. browser_take_screenshot → { type: "png", filename: ".design/onboarding-flow/screenshots/review-homepage-desktop-1280.png", fullPage: true }
    4. browser_resize   → { width: 768, height: 1024 }
    5. browser_take_screenshot → { type: "png", filename: ".design/onboarding-flow/screenshots/review-homepage-tablet-768.png", fullPage: true }
    6. browser_resize   → { width: 375, height: 812 }
    7. browser_take_screenshot → { type: "png", filename: ".design/onboarding-flow/screenshots/review-homepage-mobile-375.png", fullPage: true }

    c. Capture interactive states (when relevant).

    • Hover states on buttons, cards, links
    • Focus states on form fields
    • Open states on dropdowns, modals, menus
    • Error/success states on forms
    • Loading and empty states

    d. Capture dark mode (if the project supports it). Toggle dark mode and repeat the responsive breakpoint captures with -dark-mode in the filename.

    e. Capture specific components. If the review focuses on a particular component, use the element and ref parameters to screenshot just that element.

    Analyze Every Screenshot

    After capturing, visually analyze each screenshot against the design brief. For each screenshot:

    • Compare against the brief's aesthetic direction
    • Check visual hierarchy: is the most important element the most prominent?
    • Check spacing consistency: do margins and padding look even and intentional?
    • Check color: does the palette match the brief's direction?
    • Check typography: are font sizes, weights, and spacing visually correct?
    • Check responsive adaptation: does the layout properly reorganize (not just shrink)?
    • Note rendering issues that code review alone would miss (font loading failures, broken images, layout overflow, z-index problems, incorrect border-radius, color mismatches)

    Reference specific screenshots by filename in the review output so findings are traceable.

  4. Run the review checklist below. For each category, note what passes and what needs refinement. Be specific. Reference exact components, files, line numbers, and screenshot filenames.

  5. Produce a prioritized refinement list. Group issues by severity:

    • Must fix: Broken functionality, accessibility failures, major deviations from the brief.
    • Should fix: Inconsistencies, missing states, responsive issues.
    • Could improve: Polish, animation refinement, typography fine-tuning.
  6. Save the review as DESIGN_REVIEW.md inside the feature's .design/<feature-slug>/ folder (next to DESIGN_BRIEF.md). If no .design/ folder exists, save to the project root. Include a "Screenshots Captured" section listing all screenshots taken with their paths. Present the review directly as well if the user prefers.

Review Checklist

Visual Hierarchy

  • Is the most important content the most visually prominent on each page/view?
  • Does the type scale create clear levels of importance (heading, subheading, body, caption)?
  • Do interactive elements (buttons, links, inputs) have enough visual weight to be found without hunting?
  • Is there a clear reading order? Can you trace where the eye goes first, second, third?

Consistency

  • Are spacing values consistent? Check padding and margins against the established scale (4px/8px base or whatever the project uses).
  • Are colors used consistently? Check that the same semantic meaning always maps to the same color (primary actions, errors, success states, disabled states).
  • Are border radii, shadow values, and font sizes reused from a shared set, or are there one-off values?
  • Do similar components look and behave similarly? (e.g., all cards, all form fields, all buttons within a category.)

Aesthetic Fidelity

  • Does the implementation match the named philosophy from the brief?
  • Would someone looking at this immediately recognize the intended aesthetic direction?
  • Are there elements that break the aesthetic (a generic component in an otherwise distinctive interface, a conflicting font, an out-of-place color)?
  • Does the level of detail match the philosophy? (Minimalist designs should not have unnecessary decoration. Maximalist designs should not have empty, unfinished areas.)

Component Quality

  • Do existing components from the codebase appear correctly, or were they reimplemented?
  • Are new components following the same API patterns (props, naming, file organization) as existing ones?
  • Are there duplicate components that should be consolidated?

States and Interactions

  • Do interactive elements have all necessary states: default, hover, focus, active, disabled?
  • Do form fields have states for: empty, filled, error, success, disabled?
  • Are loading states handled? Empty states?
  • Do transitions and animations match the philosophy's motion guidelines?
  • Is there visual feedback for every user action?

Responsive Behavior

  • Does the layout work at mobile (375px), tablet (768px), and desktop (1280px+)?
  • Do components adapt appropriately? (Not just shrink, but reorganize when needed.)
  • Is touch target size adequate on mobile (minimum 44x44px)?
  • Does text remain readable at all breakpoints? No text too small, no lines too wide (max 65-75 characters).

Accessibility

  • Color contrast: Do text/background combinations meet WCAG AA (4.5:1 for body text, 3:1 for large text)?
  • Keyboard navigation: Can every interactive element be reached and activated with keyboard alone?
  • Focus indicators: Are focus rings visible and styled consistently?
  • Semantic HTML: Are headings in order? Are landmarks used (main, nav, header, footer)? Are form labels associated?
  • Screen reader: Do images have alt text? Do icons have labels? Are decorative elements hidden from assistive technology?
  • Motion: Is there a reduced-motion media query for users who need it?

Typography

  • Is the font actually loading? (Check for FOIT/FOUT flash.)
  • Are line lengths comfortable for reading (45-75 characters on body text)?
  • Is line height appropriate (1.4-1.6 for body, tighter for headings)?
  • Is the type scale intentional, or are there arbitrary sizes?

Dark Mode

  • If the project has dark mode tokens, are they applied correctly?
  • Are all color values using CSS variables (not hardcoded hex values that won't switch)?
  • Does the dark palette feel intentional for the chosen philosophy, or is it a simple inversion?
  • Are shadows adjusted for dark mode (darker, more transparent)?
  • Do accent colors maintain sufficient contrast against dark backgrounds?
  • Is there a working toggle mechanism or prefers-color-scheme support?

Mobile-First

  • Was the layout built mobile-first (using min-width media queries, not max-width)?
  • Does the mobile layout work at 375px without horizontal scrolling?
  • Is navigation adapted for mobile (not just a desktop nav that overflows)?
  • Are touch targets at least 44x44px?
  • Is body text at least 16px on mobile?

Output Format

markdown
# Design Review: [Feature/Page Name]

Reviewed against: DESIGN_BRIEF.md
Philosophy: [named philosophy]
Date: [date]

## Screenshots Captured

| Screenshot                                   | Breakpoint         | Description     |
| -------------------------------------------- | ------------------ | --------------- |
| `screenshots/review-[page]-desktop-1280.png` | Desktop (1280×800) | [what it shows] |
| `screenshots/review-[page]-tablet-768.png`   | Tablet (768×1024)  | [what it shows] |
| `screenshots/review-[page]-mobile-375.png`   | Mobile (375×812)   | [what it shows] |

> All screenshots are in `.design/<feature-slug>/screenshots/`.

## Summary

[2-3 sentences on overall quality and the biggest finding.]

## Must Fix

1. **[Issue]**: [Specific description with file/component reference]. See [`screenshots/[relevant-screenshot].png`]. _Fix: [concrete suggestion]._

## Should Fix

1. **[Issue]**: [Description]. See [`screenshots/[relevant-screenshot].png`]. _Fix: [suggestion]._

## Could Improve

1. **[Issue]**: [Description]. _Suggestion: [idea]._

## What Works Well

[Note the strongest aspects of the implementation. This is not padding. Designers need to know what to keep doing.]

Frequently asked questions

What does the Design Review AI skill do?

Run a structured design critique against the brief and codebase. Checks visual hierarchy, consistency, responsiveness, accessibility, and aesthetic fidelity. Use when user wants a design review, critique, QA pass, polish pass, or mentions "review" after building.

Why use Design Review on TypingMind?

Because you install it once and use it with any model. Design Review is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Design Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/julianoczkowski/designer-skills/tree/main/design-review. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Design Review?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Design Review?

As many as you like. As long as a model supports skills, you can use Design Review with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Design Review AI skill free?

Yes. It is published on GitHub by julianoczkowski under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇