Ai Ds Composer logo

Ai Ds Composer

Community
cris-achiardi
ai-ds-composer

Strategic guide for consuming design system metadata to compose components intelligently. Use when generating or modifying UI components and layouts to ensure component reuse, proper variant selection, and adherence to design system patterns. Helps Claude understand project structure, read hierarchical metadata indexes, apply anti-pattern rules, and flag when existing components don't fit a requirement.

Overview

Publishercris-achiardi
Repositoryclaude-skills
Skill nameai-ds-composer
Stars
72
Forks
8
Bundled files
4
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.

  • 4 bundled files

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

  • Open source

    Published by cris-achiardi on GitHub. Read the source before you install it.

Installation

Install the Ai Ds Composer 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/cris-achiardi/claude-skills/tree/main/skills/ai-ds-composer
  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/cris-achiardi/claude-skills.git /tmp/claude-skills
mkdir -p .claude/skills
cp -r /tmp/claude-skills/skills/ai-ds-composer .claude/skills/ai-ds-composer
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ai Ds Composer 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 Ai Ds Composer 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 Ai Ds Composer 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.

Design System Component Reasoning

Your role is to help generate UI using existing design system components—not to invent new code. This skill teaches you to reason through design system metadata hierarchically and make component decisions that prioritize reuse.

Core Principle

Consume metadata → Understand intent → Select components → Compose patterns → Flag gaps

Success means high component reuse with intentional, justified decisions.

Getting Started: Understanding the Project

When you encounter a design system request, your first step is understanding the project structure. Read the available documentation:

  1. Top-level system metadata - Usually in src/ or project root, shows framework decisions, philosophies, and metadata hierarchy
  2. Package.json - Reveals framework (Astro, Next.js, etc.), dependencies (React, Vue), styling approach, icon library
  3. CSS metadata (if exists) - How the project structures styling: tokens, utilities, layout patterns

If none of these exist, infer from folder structure and ask the user for clarity. The project stack directly influences which components to suggest (Astro components ≠ React patterns).

Example Metadata Hierarchy

project-root/
├── src/
│   ├── design-system.metadata.ts (top-level: project philosophy)
│   ├── styles/
│   │   └── index.metadata.ts (CSS architecture, token structure)
│   └── components/
│       ├── atoms/
│       │   ├── index.metadata.ts (all atoms, selection guide, priorities)
│       │   ├── Button/
│       │   │   ├── Button.astro (actual component)
│       │   │   └── Button.metadata.ts (variants, props, use cases)
│       │   ├── Heading/
│       │   │   ├── Heading.astro
│       │   │   └── Heading.metadata.ts
│       │   └── ... (other atoms)
│       ├── molecules/
│       │   ├── index.metadata.ts (compositions, common patterns)
│       │   └── ... (cards, form elements, etc.)
│       └── organisms/
│           ├── index.metadata.ts
│           └── ... (hero, sections, etc.)

When working with a project, you'll read metadata at multiple levels—start broad, narrow down to specific.

The Component Selection Algorithm

When you need to suggest a component, follow this flow:

1. Parse the User's Intent

Convert vague requests into clear requirements:

Vague: "Add a button like the one on the homepage" Clear: "Need a primary call-to-action button matching the hero section style, size=lg, with an arrow icon"

Tools for clarity: Ask follow-up questions, reference existing usage, look for patterns in the codebase.

2. Consult the Metadata Hierarchy (Hybrid Approach)

First pass (cache once per conversation):

  • Read design-system.metadata.ts (or infer from structure)
  • Understand project philosophies: "Prefer native HTML" vs "Use components"
  • Note the metadata organization

Subsequent passes (per component decision):

  • Read the appropriate layer index (atoms/index.metadata.ts etc.)
  • Scan the SelectionGuide for your use case
  • Pull the specific component metadata

Example workflow for "I need a button":

1. Cache: Read design-system.metadata → learn "prefer CSS over Spacer, components for interactivity"
2. Lookup: Check atoms/index.metadata → Button is high priority, use cases include "primary-call-to-action"
3. Read: Pull Button.metadata.ts → find variant options (primary/secondary/ghost/danger)
4. Decide: Based on intent, suggest variant + props
5. Verify: Check antiPatterns in Button.metadata → "Never use Button for plain navigation"

3. Apply Selection Criteria

Component metadata contains selectionCriteria or SelectionGuide. Use these to justify your choice:

typescript
// From Button.metadata.ts - these are decision rules
selectionCriteria: {
  usePrimary: "Main action user should take on the page/section",
  useSecondary: "Alternative actions, cancel buttons",
  useGhost: "Tertiary actions, minimal visual weight",
  useDanger: "Delete, remove, destructive actions"
}

Match user intent to these criteria, not your intuition.

4. Check Anti-Patterns (Hard Rules)

Anti-patterns are violations of design system philosophy. They exist in metadata as warnings.

From Button metadata:

antiPattern: "Using button for plain navigation"
reason: "Buttons indicate actions, not navigation"
alternative: "Use Link component for navigation"

Core rule: Never suggest an anti-pattern without explicit user override and human acknowledgment. If the user says "I know, but I need it anyway," that's a decision they own—but flag it clearly.

5. Suggest Composition & Variants

Once you've selected the component, determine:

  • Which component? (e.g., Button)
  • Which variant? (e.g., variant="primary")
  • What props? (e.g., size="lg", disabled={false})
  • How composed? (e.g., Button with Icon inside, following composition rules from metadata)
  • Why this choice? (cite the metadata criterion that led here)

6. Explain Your Reasoning

Always show your thinking by referencing metadata:

Good:

"You need a primary call-to-action. Button metadata shows variant="primary" is used for 'main call-to-action, high visual prominence.' Size should be lg for hero sections, following the composition pattern shown in selectionGuide.actions."

Avoid:

"You need a button. I'll use primary variant because it looks good."

When Components Don't Fit

If the user's request doesn't map to existing components, flag it explicitly:

  1. Identify the gap: "You need X, but the design system has Y. These don't align because..."
  2. Explain why: Reference what exists and why it doesn't work
  3. Don't auto-generate: Don't invent a new component. Instead:
    • Suggest a workaround (compose existing components differently)
    • Explain the implication (e.g., "This would require custom CSS")
    • Let the human decide if it warrants a new component

Example:

"Your use case needs a collapsible accordion with nested groups. The design system has no Accordion component. You could:

  • Compose with existing Button + disclosure patterns (simpler, limits nesting)
  • Create a new Accordion component (more work, but more flexible)

Which approach works for your requirement?"

Handling Vague Metadata Hints

Metadata includes AI-specific fields like aiHints and keywords. Use these as starting points, but rely on structured fields (props, variants, selectionCriteria, antiPatterns) for real decisions.

typescript
// Structured - use this for decisions
variants: { 
  variant: { options: ["primary", "secondary"], default: "primary" }
}

// Helpful but secondary - use for exploration
aiHints: { 
  keywords: ["button", "cta", "submit"]
}

CSS vs Components Decision

Your project likely emphasizes "use native HTML + CSS utilities" when possible, reserving components for complex interactive patterns.

Pattern from atoms/index.metadata.ts:

preferNativeHTML: "Use native <h1>-<h6> instead of Heading when possible"
preferCSS: "Use CSS gap in flex/grid instead of Spacer"

Apply this philosophy:

  • Layout problem? Suggest CSS utilities (grid, flexbox, gap)
  • Spacing issue? Suggest margin/padding utilities, not Spacer component
  • Simple text? Suggest native <p>, not Text component
  • Complex interaction? That's when components shine

Output Format

When suggesting components, structure your response as:

**Component:** Button
**Variant:** primary
**Size:** lg
**Props:** { href: "/features", target: "_blank", rel: "noopener noreferrer" }

**Composition:**
<Button variant="primary" size="lg" href="/features">
  Get Started
  <Icon name="ArrowRight" size={16} />
</Button>

**Why this choice:**
- Selection criteria (from Button.metadata): "Use primary for main call-to-action"
- Pattern reference: Similar to homepage hero button in ThoughtsSection
- Variant logic: lg size for hero context (from size guide)

**Notes:**
- Icon added following Button composition pattern
- External link uses rel="noopener noreferrer" (security best practice)

Working with Reference Files

When the SKILL.md body isn't enough, see the bundled references:

  • COMPONENT_SELECTION.md - Detailed component selection examples across different project types
  • METADATA_STRUCTURE.md - How to read and interpret different metadata formats
  • COMPOSITION_PATTERNS.md - Common component compositions and pattern rules
  • ANTI_PATTERNS.md - Full anti-pattern reference with explanations

Load these when you encounter complexity or need validation.

Key Reminders

  1. Metadata is authority - Your suggestions should always map back to documented metadata, not assumptions
  2. Prefer reuse - Composing existing components > generating new code (every time)
  3. Make trade-offs explicit - If you're choosing component A over B, explain why
  4. Flag gaps early - The human decides if a gap warrants a new component; you identify it
  5. Respect anti-patterns - They exist for reasons in the metadata; violations need explicit override
  6. Cite your sources - Reference which metadata file informed your decision

Next Steps

When ready to apply this skill:

  1. Understand the project structure (folder layout, package.json, top-level metadata)
  2. Parse the user's intent clearly
  3. Navigate the metadata hierarchy
  4. Apply selection criteria and check anti-patterns
  5. Suggest components with reasoning
  6. Flag gaps without auto-generating
  7. Iterate based on feedback

The goal isn't perfect UI from the first prompt—it's intelligent component reuse with clear, justifiable decisions.

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 Ai Ds Composer AI skill do?

Strategic guide for consuming design system metadata to compose components intelligently. Use when generating or modifying UI components and layouts to ensure component reuse, proper variant selection, and adherence to design system patterns. Helps Claude understand project structure, read hierarchical metadata indexes, apply anti-pattern rules, and flag when existing components don't fit a requirement.

Why use Ai Ds Composer on TypingMind?

Because you install it once and use it with any model. Ai Ds Composer 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 Ai Ds Composer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/cris-achiardi/claude-skills/tree/main/skills/ai-ds-composer. 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 Ai Ds Composer?

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 Ai Ds Composer?

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

Is the Ai Ds Composer AI skill free?

Yes. It is published on GitHub by cris-achiardi 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 👇