Design · Reference Map

A Map of Design Languages

Established movements and digital-era looks — each one a potential house-style skill, each with a place to go look at it.


Why "make a webpage" keeps producing the same look. Left unguided, a model regresses to the mean of its training data and most web UI it has seen is component-framework output (Material, Bootstrap, Tailwind defaults). A house-style skill is the corrective as it forces one opinionated canon instead of the statistical center. Any style below can become a skill the same way the editorial one did.

Legend: = deep canon (a real historical movement) · trend = recent / faddish (use with awareness). The Example column links to a visual reference — Dribbble searches for browsing community work, or an authoritative exemplar site for the canon entries.

A

Deep canon — historical movements

StyleSignatureGood forExample
Swiss / International1950s CH · Müller-Brockmann, Ruder Strict grid, sans-serif, asymmetric, generous white, objective Reports, posters, identity, wayfinding Dribbble ↗
Bauhaus1919–33 DE · Gropius, Bayer Primary red/blue/yellow, geometric shapes, functional sans Bold branding, educational, posters Bauhaus Dessau ↗
Constructivism1920s USSR · Rodchenko, Lissitzky Hard diagonals, red + black, photomontage, propaganda energy High-impact posters, manifestos, covers Dribbble ↗
De Stijl / Mondrian1917 NL · Mondrian, van Doesburg Black grid lines, primary color blocks, pure rectilinear Striking minimal layouts, art-led pages De Stijl ↗
Art Deco1920s–30s · Cassandre Geometric luxury, gold/black, strong verticals, symmetry Premium one-pagers, invitations, menus Dribbble ↗
Memphis1981 Milan · Sottsass Squiggles, confetti shapes, clashing brights, postmodern play Youthful, fun, creative brands Dribbble ↗
Editorial / Magazineyou have this skill Humanist serif, warm paper, hairlines, the measure Reading docs, guides, playbooks Practical Typography ↗
B

Digital-era looks

StyleSignatureGood forExample
Neo-brutalism / Data brutalismweb ~2014+ Raw, monospace, hard borders, high contrast, no soft shadows Dashboards, dev tools, "anti-corporate" Dribbble ↗
-ish Terminal / TUICLI heritage Monospace, green/amber on black, ASCII rules, grid-of-text Capture logs, status pages, dev/network docs The Monospace Web ↗
Material DesignGoogle 2014 Elevation/shadows, cards, bold color, motion Apps, dashboards (the "default" look) Material 3 ↗
Flat / Flat 2.0MS Metro 2010, iOS 7 2013 No (or subtle) shadows, solid bright color, simple icons Clean app UI, the post-skeuomorph reaction Flat design ↗
Glassmorphism trend~2020 Frosted blur, translucency over a vivid background Hero panels, cards — accessibility-fussy Dribbble ↗
Neumorphism trend~2019 Soft extruded monochrome shadows, low contrast Novelty controls — use sparingly neumorphism.io ↗
Skeuomorphismpre-2013 iOS Real-world textures, depth, mimics physical objects Retro / novelty interfaces Dribbble ↗
C

Mood / aesthetic & analytical

StyleSignatureGood forExample
Minimalism (Dieter Rams)"Less, but better" Whitespace, restraint, function over ornament Product pages, calm docs (the Apple register) Vitsœ · Rams ↗
Synthwave / Vaporwave / Outrun2010s referencing 1980s Neon, dark, 80s grids, chrome, sunset gradients Gaming, music, bold tech landing pages Dribbble ↗
Maximalismcounter to minimalism Dense, layered, ornamental, more-is-more Editorial fashion, zines, expressive brands Dribbble ↗
Tufte / analytical designEdward Tufte, 1983 Maximize data-ink, no chartjunk, honest marks The right way to do data & dashboards Tufte CSS ↗
D

Decompose a look into a skill — with one prompt

You don't have to build any of this by hand. Point a capable agent at an example — a Dribbble shot, a screenshot, or a live URL — with the prompt below, and it pulls the look apart and hands back a ready-to-save skill. In practice it gets you 80–100% of the way; the rest is a quick refine pass. Everything here is harness-agnostic: the output drops into a Claude Code skill, a rules file, or a saved system prompt equally well.

The decomposition prompt

Copy this, paste it into your agent, and attach or link the design you want to reverse-engineer.

You are a design-systems analyst. I'll give you a design reference —
a screenshot, an image, or a URL (for example a Dribbble shot).
Decompose it into a reusable "house-style" skill.

STEP 1 — Extract the canon (the recurring, defining traits):
• Palette   — every distinct colour as a hex value, named by role
              (bg, ink, accent, muted, line…); include any variants.
• Type      — font families by role (display / body / mono), their
              weights, the size scale, line-height and letter-spacing.
• Layout    — content width and line length (the measure), spacing
              scale, grid, border weight, corner radius, elevation.
• Signature — the 3–5 traits that make the look recognisable, plus
              the anti-patterns that would break it.

STEP 2 — Emit the skill:
1. name        — a short kebab-case identifier.
2. description — one or two sentences: what it makes, when to use it.
3. Design system — the tokens and rules above as concrete, non-
                   negotiable values (CSS custom properties for colour;
                   exact px / ch / rem for everything else).
4. Template    — one complete, self-contained HTML document that
                 already renders the look, ready to clone and refill.

Output as a SKILL.md file with YAML frontmatter (or a system prompt /
rules file if I say so). Keep every value specific; where the
reference is ambiguous, choose the most defensible option and note
the assumption.

Using it

1 · Aim it at a specimen. Paste the prompt, then attach a screenshot of the example, drop in its URL, or open the Example link from the map above. The sharper the reference, the sharper the extraction — a clean full-page shot beats a thumbnail.

2 · Save what it returns. Put the skill where your agent reads reusable instructions — a SKILL.md for Claude Code, a rules file for Cursor or Windsurf, or a saved system prompt for a bare API or chat UI. The contents are identical; only the container changes.

3 · Generate, compare, refine. Ask for a real document in the new style and hold it against the original. Feed back any drift in plain language — "accent should be #15706a, cap body text at 66ch, headings in Fraunces 900" — and regenerate. That follow-up is your second prompt; one or two passes closes the last 20%.

View on GitHub

Know a style worth adding, or spot something to fix? Open an issue → — improvements welcome.

Curated reference · fieldnotes house style, soft-aurora ground · Example links open externally (Dribbble searches show community work; canon entries link to authoritative exemplars)