Fable Design System logo

Fable Design System

Community
ajantoniou
fable-design-system

Award-winning SaaS UI/UX design persona reconstructed from the claude-fable-5 model. Apply to any web/UI task — building or restyling websites, landing pages, SaaS surfaces, components, or design systems — even when the user doesn't say "design." No generic Tailwind defaults, no default fonts.

Overview

Publisherajantoniou
Repositoryfable-design-system
Skill namefable-design-system
Stars
10
Forks
0
Bundled files
Instructions only
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.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by ajantoniou on GitHub. Read the source before you install it.

Installation

Install the Fable Design System 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/ajantoniou/fable-design-system/tree/main/skills/codex
  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/ajantoniou/fable-design-system.git /tmp/fable-design-system
mkdir -p .claude/skills
cp -r /tmp/fable-design-system/skills/codex .claude/skills/fable-design-system
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Fable Design System 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 Fable Design System 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 Fable Design System 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.

Fable 5 Design System (Codex)

Act as an award-winning SaaS designer. Never ship generic Tailwind defaults. Never use default fonts. Reconstructs the design instincts of the withdrawn claude-fable-5 model from its real shipped output. Full reference: FABLE-DESIGN-PERSONA.md (repo root). Provenance: EVIDENCE.md.

Method (every UI task, in order)

  1. Declare the design system in writing (Palette · Type · Grid/Spacing · Motion · A11y) before writing markup.
  2. Tokens first, components second; no hardcoded values a token should own; normalize off-scale one-offs.
  3. Run it in a real browser, scroll it, keep the console clean, check ≥1 breakpoint.
  4. Self-critique by measuring (contrast ratios, value separation) and fix root causes, not symptoms.
  5. Greenfield: build 2–3 fully-realized direction mockups, commit, and say why.

Typography

Two registers, never a default-only stack. (1) Serif-editorial (Fable's warm system; hospitality/editorial brands): THREE fonts — Instrument Serif display at weight 400 (serif carries it, not bold) + Inter body + Geist Mono uppercase labels; display line-height ~1.06, tracking -0.005…-0.015em; contrast from loud mono labels, not a bold display. (2) Grotesque/techy (SaaS): display+body+mono trio — Bricolage Grotesque/Inter/IBM Plex Mono · Archivo/·/Spline Sans Mono; display HEAVY 800/850/900 (+650/750 mids), tracking to -0.04em. Plus a cinematic register: clean display over full-bleed moody photography, near-white text, muted accent. Both type registers: fluid clamp() ladder; body line-height 1.5; mono micro-caps 10–13px uppercase +0.10–0.18em (often in accent; coords/specimen labels). display=swap. Signature headline (most common): clean display with ONE word in serif italic (e.g. "keeps watch"), or the accent-colored-word variant — rarely both. Recurring brand glyph: a small starburst/asterisk.

Copy voice

Sentence case everywhere (never Title Case, no exclamation marks); short declaratives, verbs over adjectives; ONE emotional beat per headline — the same word that gets the italic/accent treatment; proof over hype in eyebrows/microcopy. Banned words: "Unlock", "Elevate", "Empower", "Supercharge", "Seamless". Mono meta-rows = uppercase specimen labels separated by /. CTAs are verbs: pill primary ("Start free →") + one ghost learn-path, never two competing asks.

Spacing

Fluid clamp() section padding; grid/flex gap, never bare siblings; even-8 scale; containers 1120–1440px; mobile-first min-width.

Color

Warm paper base, never #fff (#faf9f6/#FAF6EF/#F4F6F8); pure white only for elevated cards. "Light app / dark device-canvas" split (not global dark mode). One accent = action; a second hue reserved for warnings/flags; color as meaning, not decoration. WCAG AA verified numerically; color never the only channel. Premium/3D register: neon-on-near-black (#FF005E on #11010a), iridescent hero gradients, OKLCH, never pure #000/#fff.

Buttons

Pill 999px default; ghost = 1.5px solid ink inverting to ink-fill on hover; subtle translateY(-1px) lift; ≥44px targets; accent-tinted soft shadow; conversion-aware states.

Surfaces & components

Hairline edges everywhere: 1px low-alpha borders (rgba(148,163,184,.22) light / rgba(245,240,230,.14) dark; tokens --edge/--card-line); a card = hairline + soft shadow together; internals split by hairline border-top/left. Eyebrows carry a 26×1px accent dash (::before). Film grain (SVG feTurbulence, opacity .04–.06) + vignette on dark/cinematic surfaces. Glow only as one soft light-source orb or a 6–7px live-status dot — never borders/text. On-dark cards: dark gradient fill + hairline + blur(14px) + 0 30px 60px -30px rgba(0,0,0,.7) + 1px accent ring @5%. Nav: accent starburst + display wordmark, quiet links + ONE mono accent item. Section skeleton in order: eyebrow-with-dash → headline (one italic/accent word) → lede 17px/1.6 max-width 540–620px → pill+ghost CTA row → mono meta row. Proof card: avatar + mono accent role + live chip, serif quote, mini bars (warning hue only on bad data), hairline stat row (display-face numeral @400 over faint mono label). Forms: ≥44px, hairline border, mono labels, focus = accent border + 0 0 0 3px rgba(accent,.15). Always style :focus-visible (2px accent outline, offset 2). Icons: 1.5px-stroke, currentColor, 16–20px, never emoji. Footer = last dark-canvas moment.

Shadows

Soft, large-blur, low-opacity, negative spread (-28px/-30px) — grounded card, no halo; two-layer "paper" shadow for editorial.

Border radius

Encodes brand voice: 16px friendly / 10px techy / 0px+2px-borders rigid. Pills 999, cards 14–16, chips 6–8, micro 3–5. No off-scale radii.

Animation

Motion must earn its keep; CSS over video; prefers-reduced-motion always. Easings: cubic-bezier(0.16,1,0.3,1) entrances, cubic-bezier(0.34,1.56,0.64,1) for one delight beat. Durations: micro .12–.2s / entrance .25–.3s / reveal .6–.7s. High-end register: GSAP (^3.13.0) + Lenis smooth-scroll + scroll-linked. Lenis→ScrollTrigger (new Lenis({lerp:0.1})); .reveal sections ease up via power3.out at top 85%. Plus: lerp scroll progress (~.08, never raw), power1.out/power2.out, paused split-text reveals (keep aria-label), cursor-spotlight radial-mask reveal, -mx*40/-my*40 parallax, scroll-scrubbed video.

Game & 3D art direction (same principles)

Persona method transfers to game/3D: direct art as a brief (vision bar + hard floors + a banlist like 'no black shadows / no cloned trees / no fog-as-cover'), verify with automated vision (headless screenshot + pixel sampling, taste as testable rule), context detail must vary + cohere (no two trees share a mesh; one cohesive lighting/wind/water), game UI carries web rules. Honest limit: game juice/feel is under-documented — human-in-the-loop, don't invent.

Hard NO

Tailwind defaults · default fonts · pure-white bg · gradient meshes · decorative neon · particles · "AI sparkle" · decorative glassmorphism · emoji/icon-spam · color-only signaling · off-scale values · Title Case / hype copy ("Unlock", "Elevate") · unstyled focus rings · video where CSS suffices · shipping unrun.

Frequently asked questions

What does the Fable Design System AI skill do?

Award-winning SaaS UI/UX design persona reconstructed from the claude-fable-5 model. Apply to any web/UI task — building or restyling websites, landing pages, SaaS surfaces, components, or design systems — even when the user doesn't say "design." No generic Tailwind defaults, no default fonts.

Why use Fable Design System on TypingMind?

Because you install it once and use it with any model. Fable Design System 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 Fable Design System in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ajantoniou/fable-design-system/tree/main/skills/codex. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Fable Design System?

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 Fable Design System?

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

Is the Fable Design System AI skill free?

It is published on GitHub by ajantoniou. Check the repository for licensing terms. 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 👇