agentleFS
gnegreiros-uxgnegreiros-ux/agentica-design-system/.agents/skills/ai-ds-composer/SKILL.md

ai-ds-composer

Skill2 starsChanged 46 days agoApache-2.0View on GitHub

Assemble Agentica interface patterns from a natural-language request, guaranteeing the result respects the system's component rules, tokens, and contracts — never inventing a component, a variant, or a token. Use when asked to compose, assemble, or build a UI/interface/screen from a description using Agentica components.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before returning the result, verify:

    line 41
  2. Distinctive:

    If the request requires something that does not exist in the system:

    line 53
  3. Distinctive:

    Do not invent a component

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: ai-ds-composer

Cross-tool port of .claude/skills/ai-ds-composer/SKILL.md (Claude Code's native implementation) — see governance/ai-skills-reference.md for the full picture of what Agentica's skills automate. This file follows the open Agent Skills format (agentskills.io) so Codex CLI and GitHub Copilot CLI load it automatically from .agents/skills/; content is identical in substance to the Claude Code version.

Objective

Enable interface composition in natural language while guaranteeing that the result respects the system's rules, tokens, and contracts.

Composition process

Step 1 — Understand the request

Extract from the request:

  • The user's goal (what they want to accomplish)
  • The context (page type, flow, platform)
  • The specified constraints (accessibility, language, density)

Step 2 — Read the catalog

Consult guidelines/components/overview.md to identify:

  • Available components
  • Existing patterns that already solve this need
  • Anti-patterns to avoid

Step 3 — Compose

Assemble the components while respecting:

  • Each component's rules (governance/rules/components/)
  • Visual hierarchy (only one primary per section)
  • Accessibility constraints (WCAG 2.1 AA)
  • Token dependencies

Step 4 — Validate

Before returning the result, verify:

✅ All components used exist in the system
✅ No invented variant
✅ Referenced tokens exist in component.json
✅ No hardcoded value
✅ Accessibility respected
✅ Each component's rules respected

Step 5 — Flag uncovered cases

If the request requires something that does not exist in the system:

  • Do not invent a component
  • Flag the gap to the responsible designer
  • Propose the closest available alternative

Output format

## Composition: [request title]

### Components used
- `ds-button` (variant: primary) — primary action
- `ds-input` (type: text) — name entry
- `ds-badge` (variant: success) — confirmation

### Structure
[HTML code / pseudo-code with the components]

### Tokens applied
- background: `var(--agtc-component-button-primary-background)`
- spacing: `var(--agtc-semantic-space-layout-component)`

### Rules respected
- [x] Only one primary button
- [x] Focus visible on all interactive elements
- [x] Sufficient contrast

### Uncovered cases / escalation
- [ ] The date picker pattern does not exist yet → flag to the Design System Lead

What this skill does NOT do

❌ Invent components
❌ Invent tokens
❌ Bypass accessibility rules
❌ Decide editorial content (final copy, labels)
❌ Make product architecture decisions

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.