本文へ移動
cccskills
無料GitHub で公開

design-style-installer

Install a complete design-token theme into any project in one pass -- pulls a style from the Open Agent Stack design-styles collection (aurora-mesh, cirrus, liquid-glass, mono-brutalist, neo-terminal, sand-terra, tidal) or a local tokens.json, wires it into Tailwind/CSS, and repaints existing components to match.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md3.7 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Design Style Installer

Apply a full design system to a project the way a brand refresh actually works: tokens first, wiring second, components last, and nothing hardcoded that the tokens already define. Sources: the seven production themes in Open Agent Stack design-styles/ -- each ships tokens.json (W3C design-tokens), theme.css, tailwind.tokens.js, and a WCAG contrast worksheet -- or any local W3C-format tokens file.

Workflow

  1. Pick the style. If the user named one, fetch it from the open-agent-stack repo (or read the local clone). If not, show the seven options with their one-line personalities -- aurora-mesh (soft futuristic dark, gradient mesh), cirrus (light cloud/sky, frosted), liquid-glass (translucent layered), mono-brutalist (stark, type-led), neo-terminal (CRT green-on-black), sand-terra (warm editorial neutrals), tidal (deep ocean blues) -- and ask which fits the product.
  2. Read the project. Detect the stack (Tailwind version, CSS modules, styled-components, plain CSS) and inventory the current styling surface: where colors, fonts, radii, and shadows are defined today, and how many components hardcode values outside that.
  3. Install the tokens. Wire the theme by stack: Tailwind -- merge tailwind.tokens.js into theme.extend (v4: emit @theme CSS variables); otherwise -- add theme.css custom properties at :root/[data-theme]. Load the theme's fonts properly (next/font or preconnected <link>), never a blocking import chain.
  4. Repaint. Replace hardcoded values in components with token references, mapping by role, not by nearest hex: old primary -> new accent, old page background -> bg.base, old card -> surface. Where the old design had roles the theme lacks (a third accent, a warning tint), derive them from the theme's palette logic and document the additions in the tokens file -- do not freelance new colors.
  5. Verify. Run the build, then check the money paths render correctly: nav, hero, buttons in all states, forms, cards. Confirm text/background pairs still meet the contrast worksheet's AA levels after any derivations. Output design-install-report.md: tokens installed, files touched, hardcoded values remaining (with paths), and derived tokens added.

Rules

  • Tokens are the only source of truth after installation. A repaint that leaves hex codes scattered in components has not finished.
  • Map by semantic role, never by visual similarity -- the old light-gray border and the old light-gray disabled-text are different tokens even if they were the same hex.
  • Never mix themes. If the user wants aurora-mesh cards on a sand-terra page, that is a custom theme -- fork the tokens file and name it honestly.
  • Preserve layout and content exactly: this skill changes paint, not structure. Structural itches go in the report as suggestions.
  • Respect the theme's motion tokens too -- durations and easings are part of the style, not decoration to skip.
  • Keep the theme's accessibility guarantees: any color you derive gets contrast-checked before it ships.

Quick Commands

  • "Install [style] into this project" -- full workflow
  • "Show me the styles" -- step 1 gallery with personalities
  • "Just the tokens, no repaint" -- steps 1-3
  • "What's still hardcoded?" -- the drift audit from step 5

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Audit websites for accessibility issues and WCAG compliance. Use when checking accessibility, fixing a11y issues, or ensuring WCAG compliance.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Deploy a 2-layer parallel agent hierarchy for large, parallelizable work — big refactors, multi-file migrations, codebase-wide audits, bulk generation. A top-tier commander (Fable or Opus) orchestrates the swarms; the user picks a power level (Max Power / Heavy / Balanced / Economy) that sets the Opus/Sonnet/Haiku model mix per layer. Layer 1 is 3-50+ specialist agents, each with its own full context window; Layer 2 is 2+ sub-agents per member. Includes git safety, tiered sizing, a pre-deploy gate, phantom-completion checks, and multi-wave follow-up.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Use to test whether an AI agent, prompt, or skill actually works before or after it ships — build a small eval set, score runs, and catch regressions when the prompt or model changes. Trigger when the user asks whether a prompt is good, wants to compare models or versions, sees inconsistent agent output, or is about to put an agent in front of customers.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Deploys swarms of sub-agents for massive parallel data processing tasks. Unlike agent-army (which is for code changes), this is for DATA tasks -- processing 1000 documents, analyzing datasets, bulk content generation. Configurable swarm size, task distribution, result aggregation, progress tracking, and error recovery.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Designs and deploys custom agent teams for specific business workflows. Interactive discovery of business processes, then generates complete team configurations with specialized agent roles, tool access, communication protocols, and handoff rules.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Agent-to-Agent (A2A) communication protocol. Connect two or more Claude agents that pass messages, share context, delegate tasks, and collaborate. Implements structured handoffs, shared memory, and multi-agent conversations.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

OneWave-AI のスキルをすべて見る

このスキルの問題を報告する