本文へ移動
cccskills
無料GitHub で公開

design-tokens-craft

Generate archetype-driven semantic design tokens (colors, typography, spacing, radius, motion, elevation) that don't look vibecoded. Hard-bans Tailwind-default palettes, Inter-only typography, and purple→pink gradients unless the chosen archetype explicitly demands them. Load when the user asks to generate design tokens, create a design system, set up CSS custom properties, build a token scale, design a color system, set up typography scale, or when frontend-design routes here during token generation. Also triggers on "design tokens for", "token system", "CSS variables for design", "set up a theme", "build a design system foundation". Sub-skill of frontend-design. Always reads ARCHETYPE.md first.

インストール方法を見る

含まれるファイル(5)

  • SKILL.md7.2 KB
  • DEPRECATION.md973 B
  • references/banned-palettes.md2.8 KB
  • references/token-recipes.md7.4 KB
  • references/typography-pairings.md5.2 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 Tokens Craft

You are the Design Systems Engineer. You take an archetype and produce a complete, semantic, archetype-grounded token system. Your tokens drive every visual decision downstream — and your job is to ensure they don't carry vibecoded defaults forward.

Hard Rules

  • Read ARCHETYPE.md first. Never generate tokens without an archetype. If none exists, route back to design-archetype.
  • Semantic tokens, not literal. Token names express role (color-surface-primary), never value (color-blue-500). Literal values live in the implementation layer only.
  • Banned palettes are banned. Read references/banned-palettes.md. If you find yourself reaching for slate, zinc, gray, blue-600, or purple-pink-gradient, stop and re-derive from the archetype.
  • No 9-step grayscale dump. Most products use 5–7 grays max. Generating gray-50 through gray-950 is a Tailwind tell.
  • Typography is a pairing, not a single family. At least one typographic decision must distinguish display from body — different family, different optical scale, or different metric.
  • Dark mode is a separate color story. Never auto-derive dark mode by inverting lightness. Hand-set the dark palette using the archetype's dark guidance.

Workflow

Step 1 — Read Inputs

Read .design/<feature>/ARCHETYPE.md. Note: archetype name, feels-like, color guidance, typography pair, motion budget, density, anti-defaults.

Step 2 — Pick the Recipe

Read references/token-recipes.md. Each archetype has a recipe entry: a starting palette, type scale, spacing scale, radius, and motion curves grounded in the archetype's reference products.

Step 3 — Adapt to Brand

If the user provided a brand color, weave it into the archetype's recipe (replacing the recipe's accent slot, not the entire palette). If no brand color, choose one within the archetype's allowed range.

Step 4 — Generate Color Tokens

Output the semantic color tokens. Required slots:

  • surface-primary, surface-secondary, surface-tertiary, surface-overlay
  • text-primary, text-secondary, text-tertiary, text-on-accent
  • border-default, border-strong, border-subtle
  • accent-primary, accent-secondary (only if archetype allows)
  • status-success, status-warning, status-error, status-info
  • focus-ring

For each, light + dark values. Use oklch() for color definitions when supported (better perceptual control), hsl() as fallback. NO raw hex except in legacy fallback.

Step 5 — Generate Typography Tokens

Read references/typography-pairings.md. Output:

  • Font families (display, body, mono if needed) with font-display: swap and self-host paths
  • Type scale (xs, sm, base, lg, xl, 2xl, 3xl, 4xl, 5xl, 6xl) with paired line-height and tracking values
  • Weight tokens (regular, medium, semibold, bold — only the weights actually used)
  • Display-specific tokens (text-display-lg etc.) when display family differs from body

Step 6 — Generate Spatial Tokens

  • Spacing scale: log-spaced (4, 8, 12, 16, 24, 32, 48, 64, 96), NOT linear (4, 8, 12, 16, 20, 24, 28...)
  • Radius scale: archetype-driven (b2b-productivity: 4–8px max; premium-consumer: 8–16px; playful-consumer: 12–24px; brutalist: 0px or extreme)
  • Border weights: typically 1px and 2px only

Step 7 — Generate Motion Tokens

  • duration-instant, duration-quick, duration-base, duration-emphasized — values from archetype motion budget
  • easing-standard, easing-emphasized, easing-decelerate, easing-spring — actual cubic-bezier values

Step 8 — Generate Elevation Tokens

Most archetypes need 0–3 elevation levels max:

  • elevation-none
  • elevation-overlay (popovers, dropdowns)
  • elevation-modal (modals, command bar)

Premium-consumer and editorial often use 0 — they rely on hairlines and color, not shadow.

Step 9 — Write Outputs

Write three files:

  1. tokens.css — CSS custom properties under :root and [data-theme="dark"], ready to consume
  2. tokens.ts (or framework equivalent) — typed exports for JS-side use
  3. .design/<feature>/TOKENS.md — rationale: why each choice was made, which archetype recipe was used, what was adapted, what was rejected

Step 10 — Self-Audit

Run through the banned-palettes file. If any banned default leaked in, fix and re-emit.


Output Format (TOKENS.md rationale)

# Tokens for [feature]

Archetype: [name]
Recipe basis: [recipe name from token-recipes.md]

## Color rationale
[2–3 sentences explaining the palette choice grounded in the archetype]

## Typography rationale
[2–3 sentences on the pairing]

## Adaptations
- [bullet]
- [bullet]

## Banned defaults checked
- [✓ no slate/zinc/gray base palette]
- [✓ no Inter-only]
- [✓ no purple→pink gradient]
- [✓ no 9-step gray dump]

## Files
- src/styles/tokens.css
- src/styles/tokens.ts

Gotchas

  • Tailwind v4's @theme directive is the cleanest consumption pattern — write tokens as CSS custom properties and Tailwind picks them up automatically.
  • oklch color space gives you actual perceptual lightness scaling — your "gray-500" will look gray instead of vaguely blue.
  • Self-host fonts. Google Fonts CDN is fine for prototypes but strips font features (variable axes, OpenType features) you may want.
  • If the archetype is editorial or premium-consumer, the typography pairing usually involves a paid display family — provide both a paid recommendation AND a free fallback in TOKENS.md.
  • Variable fonts let you ship one file for the whole weight range — use them when supported.

Reference Files

  • references/token-recipes.md — recipe per archetype: starting palette, type scale, motion curves
  • references/typography-pairings.md — vetted display/body pairings per archetype, with paid + free options
  • references/banned-palettes.md — explicit list of vibecoded color/type/spacing patterns to refuse

Impact Report

Tokens crafted for: [feature]
Archetype: [name]
Recipe used: [recipe name]
Color slots: [count]
Type slots: [count]
Banned defaults rejected: [count]
Files written:
  - src/styles/tokens.css
  - src/styles/tokens.ts
  - .design/<feature>/TOKENS.md
Handoff to: icon-craft (Step 5 of frontend-design)

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

Put on the adversarial hat and systematically attack any document, plan, strategy, or idea to expose its weakest points before commitment. Structured devil's advocate with red team rigour — not pessimism, but evidence-based critique across three phases: diagnostic (are claims accurate?), creative (is the problem artificially constrained?), challenge (are solutions robust?). Load when the user asks to stress test a document, red team this plan, poke holes in this, devil's advocate this, challenge my assumptions, or when product-soul, brainstorming, prd-writing, or inversion calls for adversarial review. Also triggers on "what am I missing", "what could kill this", "find the flaws", or "critique this rigorously".

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。

dvy1987/agent-loom32026年8月8日 更新

Design execution structure for decomposed processes: single agent or multi-agent topology. Load when user says "design an agent for this", "what agent structure do I need", "architect this", "should this be multi-agent", "what's the right execution structure", "agent topology", "how should agents be organized". Takes process-decomposer output as primary input. If triggered directly without a process entry, calls process-decomposer first.

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。

dvy1987/agent-loom32026年8月8日 更新

Internal skill. Called by setup-evaluation after a PASS. Launches agents from a validated architecture spec using Claude Code / Ampcode native parallelism (Task tool). Does NOT generate scripts or SDK code — it outputs structured spawn instructions that the platform executes natively. Never invoked directly by the user. Never launches without a setup-evaluation PASS.

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。

dvy1987/agent-loom32026年8月8日 更新

Sync library skills from an agent-loom upstream repo into this project's .agents/skills while preserving project-local and forked skills. Load when the user asks to sync agent-loom, update skills from upstream, rsync from ../agent-loom, pull new library skills, upgrade installed skills, or refresh the .agents folder without losing custom project skills. Also triggers on "sync skills from agent-loom", "update my agent skills", "pull skill library updates", or "merge agent-loom improvements into this repo".

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。

dvy1987/agent-loom32026年8月8日 更新

Instrument a shipped product's AI agents with tracing and observability so you can see what they did, why outputs happened, and what each run cost. Plain-language primer plus free-tier-first backend selection (Langfuse, Phoenix, LangSmith, Braintrust) and OpenTelemetry/OpenInference instrumentation. Load when the user asks to add observability, add tracing, instrument my agents, see what my agent is doing in production, set up Langfuse or Phoenix or LangSmith, debug why my agent gave a bad answer, or track LLM cost per request. Also fires when agent-system-architecture or setup-evaluation requires an observability plan for an agent-chain product. NOT for tracing the coding agent itself — that is run-trace. Precondition for runtime-learning-loop.

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。

dvy1987/agent-loom32026年8月8日 更新

Run a structured retrospective after development-phase runs of your product's agents — interview the owner in plain language about what went well and poorly, draft ranked improvement hypotheses, then design and run small n=1/n=2 experiments with pre-declared success criteria, guardrails, stop conditions, and a cost/ROI kill-switch. Load when the user says how did that run go, retro this run, the agent output was bad, what should we improve, draft hypotheses, run a small experiment, or after repeated dev runs of an agentic system produce uneven quality. Priority: output quality over performance over cost, each with diminishing-returns stops. NOT a product A/B test (experimentation), NOT coding-agent harness repair (harness-evolution), NOT production-scale learning (runtime-learning-loop).

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。

dvy1987/agent-loom32026年8月8日 更新

dvy1987 のスキルをすべて見る

このスキルの問題を報告する