brand-tokens

Manzanita Research visual identity tokens — colors, typography, and texture direction. Apply these whenever building any visual interface (web, terminal, print, slides) for a Manzanita project. Other

SKILL.md

GitHub

Manzanita Research — Visual Identity Tokens

This is the single source of truth for Manzanita's visual identity. Colors, typography, and texture direction live here. Design skills (frontend-design, tui-design) reference this file for brand-specific values — the craft guidance lives in those skills, not here.

The Manzanita Look

Whole Earth Catalog meets Laurel Canyon. Analog warmth in digital spaces. Sun-faded, textured, unhurried. Like a workshop with good light and interesting tools on the walls.

What it is: Warm, editorial, handmade-feeling. Film stock over digital. Dense information presented with care. Calm confidence.

What it is not: Startup templates. Purple gradients. Glassmorphism. Neon accents. Pure black-on-white. Rainbow ANSI vomit. Anything built to impress in a demo GIF rather than to use every day.

Color Palette

These are the canonical Manzanita colors. Use the semantic names, not raw hex — it makes intent clear.

Token Hex Role
cream #FAF9F6 Page backgrounds, canvas
warm-black #2C2C2C Primary text
bark #6B3A2A Deep accent, manzanita bark. Borders, emphasis
terracotta #C2714F Primary warm accent. CTAs, active states
sage #8B9E7E Success, healthy, linked, good states
ochre #C49A3C Highlights, warnings, gold accent
rust #A0522D Tertiary accent, error states
lavender-dried #9B8EA8 Subtle accent, tags, metadata
fog #E8E5DF Borders, dividers, subtle separators
dusk #5C5C5C Muted/secondary text

Color rules

  • Use color sparingly. Most of the interface should be cream, warm-black, fog, and dusk. Color draws attention — use it where attention belongs.
  • Status uses the landscape palette: sage for good, ochre for caution, rust for problems, terracotta for active/selected. These map to natural states (green growth, golden warning, red-brown danger) without the aggressive pure-red/green/blue of most UIs.
  • Never use pure red, green, or blue. Everything is warm and muted.
  • Respect system preferences. Support prefers-color-scheme on web, NO_COLOR in terminals. The palette works on both light and dark backgrounds — on dark, swap cream and warm-black roles.

Tailwind config (web)

colors: {
  cream: '#FAF9F6',
  'warm-black': '#2C2C2C',
  bark: '#6B3A2A',
  terracotta: '#C2714F',
  sage: '#8B9E7E',
  ochre: '#C49A3C',
  rust: '#A0522D',
  'lavender-dried': '#9B8EA8',
  fog: '#E8E5DF',
  dusk: '#5C5C5C',
}

Lip Gloss constants (terminal)

const (
    colorCream      = lipgloss.Color("#FAF9F6")
    colorWarmBlack  = lipgloss.Color("#2C2C2C")
    colorBark       = lipgloss.Color("#6B3A2A")
    colorTerracotta = lipgloss.Color("#C2714F")
    colorSage       = lipgloss.Color("#8B9E7E")
    colorOchre      = lipgloss.Color("#C49A3C")
    colorRust       = lipgloss.Color("#A0522D")
    colorLavender   = lipgloss.Color("#9B8EA8")
    colorFog        = lipgloss.Color("#E8E5DF")
    colorDusk       = lipgloss.Color("#5C5C5C")
)

Typography

Three typefaces. Each has a specific job.

Display: Fraunces

Variable font from Google Fonts. Warm, expressive, slightly wonky — the brand voice in type.

Critical settings — always apply both:

font-variation-settings:
  "WONK" 1,
  "SOFT" 100;

Never use Fraunces without wonk and full soft. Without these it looks like a different font entirely.

Loading from Google Fonts (include all variable axes):

family=Fraunces:ital,opsz,wght,SOFT,WONK@0,9..144,100..900,0..100,0..1;1,9..144,100..900,0..100,0..1

Optical sizing (opsz) is automatic — don't set it manually.

Use for: Headlines, hero text, feature titles, anything that needs personality. Set large, with negative letter-spacing (-0.02em) and tight line-height (1.1).

Body: Literata

Warm, bookish variable serif. Designed for long reading — looks best at 16px. Load weights 400 and 600 from Google Fonts.

Alternatives (if Literata isn't available): Lora, Crimson Pro, Libre Baskerville. All are warm serifs that pair well with Fraunces.

Use for: Body text, documentation, long-form reading. Set at font-size: 16px with generous line-height (1.7–1.8), comfortable measure (max-width: 38rem), and text-wrap: pretty where supported.

Monospace: Commit Mono

Custom Manzanita build, self-hosted from org/fonts/CommitMono-Variable.woff2.

OpenType features — always enable:

font-feature-settings:
  "cv01" 1,
  "cv03" 1,
  "cv04" 1,
  "cv06" 1,
  "cv11" 1,
  "ss01" 1,
  "ss02" 1,
  "ss03" 1,
  "ss04" 1,
  "ss05" 1;

Use for: Code blocks, UI labels in tool interfaces, terminal output, data tables. Anywhere precision matters.

Fonts to avoid

Never use as primary typefaces: Inter, Roboto, Arial, Space Grotesk, Poppins, Geist Mono, system defaults. These are fine fonts — they're just not Manzanita. The whole point of the type system is warmth and character over geometric neutrality.

Texture & Atmosphere

The Manzanita look has physical presence. It shouldn't feel like a flat digital surface.

  • Film grain overlays: Subtle noise at 0.03–0.08 opacity. Adds warmth and breaks up flat color fields. On web, use a CSS pseudo-element or SVG filter. In image prompts, specify film stock.
  • Warm gradients: Cream to fog, terracotta to rust. Always warm-to-warm, never cool tones.
  • Material surfaces: Paper-like textures on containers. Slight off-white variation rather than perfectly uniform backgrounds.
  • Warm-tinted shadows: Never pure black shadows. Use warm-black at low opacity, or bark-tinted shadows for colored elements.
  • Photography feel: Candid, natural-light, slightly grainy. Film stock over digital. If generating images, specify: "Kodak Eastman 100T 5247 35mm film, grain, halation, faded warm colors."

When to Use This Skill

This skill provides the raw tokens. How to apply them depends on the medium:

  • Building a web interface? Use this + the frontend-design skill (React + Tailwind craft guidance)
  • Building a terminal UI? Use this + the tui-design skill (Go + Bubble Tea craft guidance)
  • Writing copy? Use the brand-voice skill instead — that's about language, not visuals
  • Generating images? Use the generate-hero-prompt skill — it has its own visual language built in
  • Just need the hex codes or font names? This skill is all you need