Frontend Design logo

Frontend Design

Community
Rylaispirit
frontend-design

Design, implement, and verify production frontend interfaces while respecting the existing repository, accessibility, responsive behavior, interaction, and visual QA.

Overview

PublisherRylaispirit
Repositoryrylai-codex-hermes-skills
Skill namefrontend-design
Stars
63
Forks
42
Bundled files
1
LicenseMIT
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.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

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

Installation

Install the Frontend Design 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/Rylaispirit/rylai-codex-hermes-skills/tree/main/skills/frontend-design
  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/Rylaispirit/rylai-codex-hermes-skills.git /tmp/rylai-codex-hermes-skills
mkdir -p .claude/skills
cp -r /tmp/rylai-codex-hermes-skills/skills/frontend-design .claude/skills/frontend-design
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Frontend Design 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 Frontend Design 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 Frontend Design 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.

Frontend Design

This skill is a practical build workflow maintained and adapted by Rylai for Codex, Hermes, and Claude. Its goal is a usable, coherent interface that survives real content, real interaction, and real viewport changes.

Runtime Contract

  • Work inside the user's existing repository when one is provided.
  • Follow higher-priority project instructions and established design tokens.
  • Reuse the installed framework, component library, and icon set when suitable.
  • Resolve files relative to this skill or the active project, never a private path from another machine.
  • Check tool and package availability before relying on them.
  • Use the runtime's browser, screenshot, and image-generation capabilities by their local names.

Required Outcome

A finished interface must:

  • perform the requested workflow rather than merely illustrate it;
  • preserve or deliberately extend the product's visual language;
  • support keyboard, pointer, touch, zoom, and reduced-motion use;
  • remain readable from narrow mobile screens through wide desktop screens;
  • expose loading, empty, error, disabled, success, and destructive states when the workflow can reach them;
  • avoid layout shifts caused by labels, icons, validation text, or late data;
  • be verified in a real browser before completion is reported.

Workflow

1. Inspect Before Designing

Read the relevant routes, components, styles, assets, package manifest, and tests. Identify:

  • the user and their main task;
  • the current navigation and information hierarchy;
  • reusable components and tokens;
  • framework and build constraints;
  • nearby screens that establish product conventions;
  • existing user changes that must remain intact.

For a reference image or existing page, separate structural facts from decorative details. Record the important relationships: hierarchy, alignment, density, rhythm, image treatment, interaction priority, and responsive intent.

2. Define the Experience

Write a short internal brief before editing:

  • Primary task: what the user must accomplish.
  • Priority content: what must be visible first.
  • Interaction path: the shortest reliable path through the task.
  • Visual character: a few concrete qualities tied to the domain.
  • Constraints: framework, data, accessibility, performance, and deadlines.
  • Proof: which interactions and viewports will demonstrate completion.

Do not add a marketing introduction when the request is for a working tool. Start with the actual product surface.

3. Plan the Screen Structure

Choose the smallest structure that supports the workflow:

  • persistent navigation only when users need repeated movement;
  • toolbars for frequent commands;
  • tabs for peer views;
  • menus for compact option sets;
  • dialogs for focused, reversible interruption;
  • cards only for genuinely repeated, self-contained items;
  • tables or aligned lists when comparison matters.

Map each important action to a visible control. Prefer familiar icons for common tool actions and provide an accessible name or tooltip when the meaning is not obvious.

4. Establish the Visual System

Derive a small set of reusable decisions:

  • typography roles for display, section, body, label, and data text;
  • spacing steps and content width;
  • surface, border, text, accent, success, warning, and danger colors;
  • corner radius and elevation rules;
  • stable sizes for controls, rows, panels, and fixed-format content;
  • focus, hover, pressed, selected, and disabled treatments.

Use color to express hierarchy and state. Do not let one hue dominate every surface. Choose imagery and illustration that reveal the actual product, subject, or task rather than acting as unrelated atmosphere.

5. Implement the Core Path

Build the semantic structure first, then styling, then motion. Keep state close to the component that owns it unless the repository has an established state pattern. Prefer existing helpers and native platform behavior over new abstractions.

The primary path must be functional. Connect real data when the project provides it; otherwise use explicit local fixtures that can be replaced without rewriting the component.

6. Complete Interaction States

For every interactive control, verify:

  • pointer and keyboard activation;
  • visible focus;
  • disabled behavior;
  • progress feedback for delayed actions;
  • validation near the field or action that caused it;
  • recovery guidance for failures;
  • confirmation for destructive or irreversible actions;
  • stable geometry while labels and messages change.

Do not leave controls that look clickable but perform no action.

7. Make Responsiveness Intentional

Design responsive behavior from content pressure, not device labels.

  • Let grids reduce columns before content becomes cramped.
  • Allow toolbars to wrap, collapse, or move secondary actions into a menu.
  • Preserve the primary action and essential context on small screens.
  • Give tables an explicit narrow-screen strategy: horizontal scroll, selected columns, stacked records, or a detail view.
  • Use minmax(), clamp() for dimensions, aspect-ratio, and bounded containers where they prevent layout drift.
  • Wrap long labels and test unusually long values.

Check at least one narrow mobile width, one tablet-like width, and one desktop width. Add a wide-screen check for dense dashboards or full-bleed media.

8. Apply Accessibility

Use semantic HTML before ARIA. Confirm:

  • one clear page heading and logical heading order;
  • labels for inputs and accessible names for icon controls;
  • keyboard traversal that follows the visible reading order;
  • focus that is never hidden by overlays or sticky regions;
  • sufficient text and control contrast;
  • status messages announced when needed;
  • dialogs that manage focus and close predictably;
  • meaningful alternative text for informative images;
  • decorative images excluded from assistive output;
  • no information conveyed by color alone;
  • motion reduced when the user requests it;
  • content remains usable at browser zoom.

Use automated checks as a supplement to keyboard and visual inspection.

9. Control Performance

  • Avoid shipping large media when a smaller representation is enough.
  • Reserve dimensions for images and asynchronous content.
  • Lazy-load below-the-fold media when appropriate.
  • Keep animation on opacity and transforms when possible.
  • Avoid unnecessary rerenders and broad dependency additions.
  • Respect the project's established bundling and data-fetching strategy.

10. Verify in a Browser

Run the project's normal checks, start the appropriate local server, and inspect the interface in a real browser.

Verification should cover:

  1. initial render and console output;
  2. the full primary interaction path;
  3. loading, empty, error, and success states;
  4. keyboard navigation and visible focus;
  5. narrow, medium, desktop, and relevant wide viewports;
  6. long content, validation messages, and overflow;
  7. screenshots for visual comparison and overlap detection;
  8. production build or equivalent repository check.

When a visual reference exists, compare screenshots against it and correct hierarchy, spacing, scale, cropping, and alignment. Do not claim fidelity from code inspection alone.

Domain Modes

Operational Products

Favor scanability, predictable navigation, compact controls, clear status, and fast repeated action. Dashboards, CRMs, editors, and admin tools should spend space on useful information rather than decorative introductions.

Content and Editorial Experiences

Prioritize reading rhythm, meaningful imagery, clear sequencing, and restrained navigation. Preserve enough visible continuation that users understand there is more content below the first viewport.

Brand and Product Pages

Make the brand, product, place, or person unmistakable in the first viewport. Use real or generated bitmap media that shows the subject clearly. Supporting copy explains value; the main heading names the offer or subject.

Games and Creative Tools

Allow stronger motion and illustration, but keep controls learnable and state legible. Use a proven engine for established game rules, physics, parsing, or 3D behavior when the project permits it.

Completion Gate

Before finishing, confirm:

  • the requested workflow works end to end;
  • no unrelated files were changed;
  • no visible element overlaps or clips at tested widths;
  • text fits controls and containers with realistic long content;
  • controls have complete states and accessible names;
  • keyboard and reduced-motion behavior are usable;
  • the browser console has no new actionable errors;
  • tests and the production build pass, or limitations are reported precisely;
  • screenshots represent the final implementation, not an earlier draft.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Frontend Design AI skill do?

Design, implement, and verify production frontend interfaces while respecting the existing repository, accessibility, responsive behavior, interaction, and visual QA.

Why use Frontend Design on TypingMind?

Because you install it once and use it with any model. Frontend Design 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 Frontend Design in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Rylaispirit/rylai-codex-hermes-skills/tree/main/skills/frontend-design. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Frontend Design?

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 Frontend Design?

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

Is the Frontend Design AI skill free?

Yes. It is published on GitHub by Rylaispirit under the MIT 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 👇