Web Design Direction System logo

Web Design Direction System

Community
TheGoat395
web-design-direction-system

Resolve shared design intent and guidance.

Overview

PublisherTheGoat395
RepositoryCodex-Skills
Skill nameweb-design-direction-system
Stars
136
Forks
48
Bundled files
2
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.

  • 2 bundled files

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

  • Open source

    Published by TheGoat395 on GitHub. Read the source before you install it.

Installation

Install the Web Design Direction System 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/TheGoat395/Codex-Skills/tree/main/skills/web-design-direction-system
  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/TheGoat395/Codex-Skills.git /tmp/Codex-Skills
mkdir -p .claude/skills
cp -r /tmp/Codex-Skills/skills/web-design-direction-system .claude/skills/web-design-direction-system
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Web Design Direction System 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 Web Design Direction System 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 Web Design Direction System 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.

Web Design Direction System

Use when shared visual intent is unresolved or a recurring design correction requires revising the governing guidance. A settled design or localized implementation goes directly to its specialist; this is not a mandatory router.

Design Guidance Contract

When a project has repeated design decisions, a named design system, or a recurring review failure, read the design guidance evidence loop before changing the standard. Start with the reader's job and the strongest supported answer, then choose the smallest destination that can enforce the decision: guidance, a component primitive, a deterministic check, an exemplar, or an evaluation case.

Do not turn one screenshot, one reviewer preference, or one shipped implementation into a universal rule. Record scope, rationale, evidence, exceptions, and a bad/good example. Keep unresolved decisions visible instead of hiding them inside CSS or component code.

Infer The Design Mode

Before choosing components or CSS, write a compact design read:

  • product type: marketing site, product UI, dashboard, portfolio, service business, commerce, docs, app shell, or campaign
  • user state: browsing, evaluating, comparing, buying, onboarding, operating, recovering, or managing
  • brand posture: quiet, precise, editorial, luxurious, playful, industrial, technical, cinematic, utilitarian, or warm
  • density need: sparse, normal, dense, or expert-dense
  • motion need: none, small tactile, section choreography, scroll scene, spatial/canvas, or demo-like

Turn those choices into concrete layout, type, color, media, and motion decisions.

Direction Dials

For an open brief, these optional internal dials can help compare directions. They are descriptive aids, not scores or required deliverables:

  • layout_variance: 1 centered conventional, 5 balanced asymmetry, 10 experimental composition
  • motion_depth: 1 static with hover, 5 section transitions, 10 scroll/canvas/spatial choreography
  • surface_density: 1 spacious editorial, 5 normal product page, 10 compact operator dashboard
  • brand_commitment: 1 neutral system UI, 5 clear accent identity, 10 brand carried by color, material, and type

Illustrative starting profiles: premium public site 6 / 5 / 4 / 6; product dashboard 4 / 3 / 7 / 3; expressive campaign 8 / 7 / 4 / 8. Derive actual choices from the brief; do not impose these values.

Design Lanes

Pick one lane and stay coherent:

  • editorial: strong type hierarchy, unusual grid, restrained UI chrome, image/copy rhythm
  • product-precise: crisp controls, dense proof, neutral palette, exact states, product credibility
  • luxury-service: quiet rhythm, tactile materials, high-quality imagery, short confident copy
  • cinematic: large spatial scenes, scroll pacing, image/video depth, minimal but strong copy
  • industrial: mechanical grid, hard contrast, exposed structure, deliberate tension
  • soft-premium: calm color, generous whitespace, rounded tactile controls, controlled spring motion

If the repo already has a system, preserve its best parts and change only what is weak.

Remove Generic Output

Rewrite these before shipping:

  • centered hero with vague headline and three feature cards
  • repeated bento or card grids without a reason for the grid
  • random gradients, blobs, dots, glows, browser chrome, decorative lines, or stock-looking dashboards
  • huge display words that do not carry meaning
  • unsupported metrics, testimonials, awards, logos, screenshots, or proof
  • body text too light to read
  • every section using the same entrance animation
  • mobile layouts that simply stack desktop sections
  • vague CTAs where the action can be specific
  • placeholder copy or placeholder code comments in final output

Redesign Protocol

For existing projects:

  1. Inspect the rendered UI or existing files before proposing direction.
  2. Preserve working IA, routes, real copy, assets, and brand anchors unless they are the issue.
  3. Identify the material failures, if any: hierarchy, spacing, type, color, states, motion, content, accessibility, or responsiveness.
  4. Change the smallest set of files that fixes the visible problem.
  5. Compare before and after behavior, not just code diffs.

Image-To-Code Protocol

When the user supplies a reference image or generated comp:

  1. Extract layout skeleton, type scale, spacing rhythm, color roles, surface materials, media crop rules, component vocabulary, and motion implications.
  2. Separate what to match from what to reinterpret.
  3. Preserve semantic HTML and responsive behavior instead of copying pixels blindly.
  4. Implement real components and states, not a screenshot imitation.
  5. Run visual QA against desktop and mobile after implementation.

Motion Rules

Motion should serve sequence, focus, depth, state, continuity, narrative, artistic expression or deliberate delight. Remove motion that lacks a purpose in the actual brief.

Prefer one memorable motion idea over many identical entrance effects. Always define reduced-motion behavior. Never hide content behind animation that may fail in a paused tab, slow browser, or headless QA run.

Evidence Loop

For a repeated design correction, freeze a representative prompt, inputs, viewport, model, and skill version. Save one baseline without the proposed guidance, then compare first-attempt treatment output under the same conditions. Review the pair against a written rubric, keep a small holdout case, and encode only corrections that generalize.

Separate retrieval from application: confirm that the right skill or reference loaded before judging whether the rule was followed. Keep deterministic checks for observable mechanics and keep product judgment in scoped prose. Feed repeated production corrections back into a pending review record rather than editing the standard automatically.

Preflight Before Handoff

Check:

  • design lane matches the brief
  • optional dials support the chosen direction without overriding the brief
  • generic output has been removed
  • typography is readable and not cramped
  • color contrast is acceptable for body, muted, and placeholder text
  • mobile composition has deliberate hierarchy
  • states exist for hover, focus, loading, error, empty, disabled where relevant
  • motion has purpose and reduced-motion fallback
  • no unsupported claims were added
  • final summary says what was changed and what was not verified

Read the evidence-loop reference when a change updates a shared design standard, evaluation protocol, or recurring correction pattern.

Scope and evidence

Own the shared direction record. Preserve explicit user direction and adequate existing context; blueprint, style director and build gate consume the same spec. Dials are optional internal tools and need not appear in a deliverable.

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 Web Design Direction System AI skill do?

Resolve shared design intent and guidance.

Why use Web Design Direction System on TypingMind?

Because you install it once and use it with any model. Web Design Direction System 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 Web Design Direction System in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/TheGoat395/Codex-Skills/tree/main/skills/web-design-direction-system. 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 Web Design Direction System?

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 Web Design Direction System?

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

Is the Web Design Direction System AI skill free?

Yes. It is published on GitHub by TheGoat395 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 👇