---
name: ui-skills-root
description: Use before UI-related work to select the smallest useful UI Skills context through the ui-skills CLI.
risk: unknown
source: https://github.com/ibelick/ui-skills/tree/main/skills/ui-skills-root
source_repo: ibelick/ui-skills
source_type: community
date_added: 2026-07-01
license: MIT
license_source: https://github.com/ibelick/ui-skills/blob/main/LICENSE
---

# UI Skills Root
## When to Use

Use this skill when you need use before UI-related work to select the smallest useful UI Skills context through the ui-skills CLI.


You are the routing layer for UI Skills.

This skill is shown by `npx ui-skills start` and is also available in the registry.

Use it when an agent in Codex, Cursor, or Claude Code has a clear UI goal.

If the goal is unclear, ask one short question.

If the goal is clear, choose the right category, load the smallest useful skill context, then implement.

## Protocol

1. decide if the task is UI-related
2. if not, return `no skill needed`
3. identify the likely category
4. inspect that category with the CLI
5. select the smallest useful skill set
6. load only selected skill(s)
7. implement using that context

## CLI

```bash
npx ui-skills start
npx ui-skills categories
npx ui-skills list --category <category>
npx ui-skills get <slug>
```

## Selection Rules

Prefer 1 skill.

Use 2 only when the task needs two clear angles.

Use 3 only for broad review, redesign, or multi-surface work.

Never use more than 3.

Route by topic, then stack, then specificity.

Prefer specific skills over broad skills.

Prefer framework-specific skills when the stack is obvious.

For quick cleanup, prefer the most specific craft, visual, or layout skill available.

If unsure, inspect categories and pick the safest narrow skill.

## Limitations

- Use this skill only when the task clearly matches its upstream source and local project context.
- Verify commands, generated code, dependencies, credentials, and external service behavior before applying changes.
- Do not treat examples as a substitute for environment-specific tests, security review, or user approval for destructive or costly actions.