Deft
Ask AI

Search

Search skills, authors and topics

fatihaydost / brand-identity-skill

brand-identity

Brand identity sets from one brief - logo, typography and colour palette built together, measured (contrast, colour blindness, glyph coverage, font licences, small-size logo renders) and compared side by side as identity cards, shown on the user's own site, then a brand guidelines kit (PDF, logo files, tokens) for the chosen set. Use for a new brand or any part of one (logo, fonts, colours, or a mix), for refreshing an identity, or for critiquing one, even if the user only says "branding".

View sourceVisualTypographyColorSystems

Install

npx skills add https://github.com/fatihaydost/brand-identity-skill --skill brand-identity

AI defaults

Bias, not bans: a ban tends to move the model to the next default (Anthropic 2025) [practice]. Data, match rules and tag vocabulary: assets/ai-defaults.json.

Reflex test

After drafting a set, ask: "Would I propose this same font, palette or mark idea for another brand in this sector?" If yes, justify it from the brief or change it. At least one set uses no listed item. A hit goes into defaults_used: [{id, why}]; a valid why is concrete: the user asked, the competitor map shows the region empty, an existing asset, a production constraint. "Feels warm / premium" is not a reason: it is how the stereotype works.

A. Model defaults (measured 2026-10-02; Claude Opus/Sonnet 5.5, Haiku 4.5)

Our experiment, 30 briefs, 180 answers (tools/ai_defaults_experiment.py). [limited] The default is a sector stereotype: the same brief got the same display face 67–80% on repeat, 60% across models.

  1. Fraunces display (and soft-serif neighbours such as Recoleta): 24% of display faces, Opus 40%.
  2. Inter text face: 40% of text faces.
  3. Cream ground + terracotta + gold/brass accent (+ serif display): strongest cluster, Opus 50%, Sonnet 47%.
  4. Navy primary + gold accent for law, property, finance, insurance, public sector.
  5. Sector fonts: Cormorant Garamond (wine, hotel, property), Canela (law), Barlow Condensed (construction, logistics), Space Grotesk / Sora (AI, crypto, fintech), Poppins (edtech).
  6. Indigo/violet in technology briefs only; "AI purple" is not a default elsewhere (6% of primaries).
  7. Logo formula: the name's object + hidden letter + negative space + circle + local "nod"; symbol + wordmark in 97–100%, so also offer a wordmark-only or typographic set.
  8. Gradient, blue to purple: rare (5%), a cliché when it comes.

B. Category clichés (human-made logos)

Four-point spark, radial swirl, gradient orb (Fast Company 2026); initial in a rounded square (that tile is the app-icon container); hexagon/nodes/cube; leaf, shield, roof, arrow, bulb, globe, bolt (Left Hand Design); the swoosh (Geismar, Print 2011); outline icons (Lucide's 2 px, round-cap grid); mockup polish over a weak form (judge flat, one colour, 16 px first). [practice] Before building, name three clichés of the category. A B-hit's why says who on the competitor shelf uses it, what second reading is this brand's, and whether the idea survives without it. Legitimate: tile as the app icon, gradient with a flat master, hexagon for "Hex…".

Tag marks honestly in identity.logo.tags; the audit only flags.