Skip to content
Back to search

brand-identity-skill

MarketplaceWorks with Claude CodeGitHubIncomplete

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/fatihaydost/brand-identity-skill with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/fatihaydost/brand-identity-skill'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/fatihaydost--brand-identity-skill/badge.svg)](/listings/fatihaydost--brand-identity-skill)

HTML

<a href="/listings/fatihaydost--brand-identity-skill"><img src="/api/listings/fatihaydost--brand-identity-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Claude Code plugin** (recommended):

/plugin marketplace add fatihaydost/brand-identity-skill
/plugin install brand-identity@brand-identity-skill

**Or as a plain skill**, from the `brand-identity.zip` of the [latest release](https://github.com/fatihaydost/brand-identity-skill/releases/latest) (unzip it into `~/.claude/skills/`), or from a clone:

git clone https://github.com/fatihaydost/brand-identity-skill
mkdir -p ~/.claude/skills
ln -s "$PWD/brand-identity-skill/skills/brand-identity" ~/.claude/skills/brand-identity

**Requirements:** Python 3.10+, a Chromium-based browser (Chrome, Chromium, Edge or Brave) and, ideally, [uv](https://docs.astral.sh/uv/). With uv the scripts install their own Python packages (fonttools, uharfbuzz, skia-pathops) into a cached environment on first use; nothing to `pip install`. Run `python3 skills/brand-identity/scripts/brand.py check` to see what is missing. Tested end to end on Linux, including a clean container; unit tests run in CI on Linux, macOS and Windows.

**Other hosts.** The skill is a standard `SKILL.md` folder, so agents that read Agent Skills can load `skills/brand-identity`; it is tested end to end in Claude Code. In claude.ai chat the code sandbox does not allow Google Fonts by default, so type measurement and rendering fail there; use Claude Code.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 31 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. {
  2. "name": "brand-identity-skill",
  3. "owner": {
  4. "name": "Fatih Aydost"
  5. },
  6. "metadata": {
  7. "description": "A brand-identity skill for Claude and Agent Skills-compatible agents.",
  8. "version": "1.1.1"
  9. },
  10. "plugins": [
  11. {
  12. "name": "brand-identity",
  13. "source": "./",
  14. "description": "Brand identity sets from one brief: logo, typography and colour palette built together, measured (contrast, colour-blindness, glyph coverage, licences, small-size logo renders), compared side by side as identity cards, shown on your own site, then a brand guidelines kit for the one you pick.",
  15. "version": "1.1.1",
  16. "category": "design",
  17. "keywords": [
  18. "brand",
  19. "identity",
  20. "logo",
  21. "typography",
  22. "fonts",
  23. "color",
  24. "palette",
  25. "brand-guidelines",
  26. "accessibility",
  27. "design-tokens"
  28. ]
  29. }
  30. ]
  31. }
README.md · README · 218 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <div align="center">
  2. <picture>
  3. <source media="(prefers-color-scheme: dark)" srcset="docs/images/logo-dark.png">
  4. <img src="docs/images/logo-light.png" width="96" alt="brand-identity logo">
  5. </picture>
  6. # brand-identity
  7. [![Tests](https://github.com/fatihaydost/brand-identity-skill/actions/workflows/test.yml/badge.svg)](https://github.com/fatihaydost/brand-identity-skill/actions/workflows/test.yml)
  8. [![Release](https://img.shields.io/github/v/release/fatihaydost/brand-identity-skill)](https://github.com/fatihaydost/brand-identity-skill/releases/latest)
  9. [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
  10. </div>
  11. A skill for Claude Code that designs a brand identity as one system: **logo, typography and colour palette built
  12. together from one idea**, measured, shown side by side as identity cards, and handed over as a brand guidelines kit
  13. for the set you pick.
  14. ![24 logos for 24 different brands, all made by the skill](docs/images/hero.jpg)
  15. <sub>Twenty-four briefs, one direction shown from each: a light festival in Lyon, a taquería in Mexico City, a
  16. pharmacy in Athens, a techno label in Berlin, a floating sauna in Oslo, a children's dentist, a weeding-robot
  17. start-up, a honey co-op on the Black Sea and more. Mascots, emblems, monograms, stencils and plain wordmarks; every
  18. mark, typeface and colour above came out of a run of this skill.</sub>
  19. ![A short animated walk-through of the skill](docs/images/brand-identity.gif)
  20. **[What you get](#what-you-get)** · **[Examples](#examples)** · **[Install](#install)** · **[Use](#use)** ·
  21. **[How it works](#how-it-works)** · **[Network and privacy](#network-and-privacy)** · **[Develop](#develop)**
  22. ## What you get
  23. - **Three (up to four) identity sets**, each built around a single mechanism that produces the mark, the type choice
  24. and the palette at once, so the parts belong together instead of being picked separately.
  25. - **Measured, not asserted.** Contrast (WCAG 2.2), colour-blindness simulation, glyph coverage for the
  26. languages the brand writes in, font licences, tabular figures, and logo renders at 16 and 32 px for review. A failed
  27. measurement is a gate the set must pass; a heuristic is labelled as a warning; taste is labelled as judgement.
  28. - **Sets that don't look like every other AI brand.** The skill knows what plain models reach for and asks for a
  29. reason from the brief whenever a set uses it ([below](#why-the-sets-dont-look-like-every-other-ai-brand)).
  30. - **A brand guidelines kit** for the chosen set: an 8-page PDF (cover, idea, logo, variations, logo on colour and
  31. misuse drawn for *this* mark, colours, typography, applications; `--full` adds grid, dark mode, accessibility,
  32. iconography and imagery for 13), logo files (SVG and PNG: full colour, one colour, reversed, app icon, favicon), design tokens (CSS,
  33. SCSS, Tailwind, DTCG JSON) and a font sheet with embed code.
  34. - **Your site, restyled.** Give it a URL and it reads the current logo, type and colours, then previews each set on
  35. your own pages. Keep the parts you like (`keep`), evolve others (`refresh`).
  36. - **Any part on its own**: just a logo, just a type system, a palette for an existing logo, or a critique of an
  37. identity you already have.
  38. ## Examples
  39. ### Fernhill Books: an independent bookshop and small press in Edinburgh
  40. *Brief: literate, unhurried, wry. Competitors: a national chain in navy and white, a second-hand shop with a gilt
  41. sign, an online marketplace.*
  42. ![Three directions for Fernhill Books with the comparison table](docs/images/fernhill-board.jpg)
  43. **A · Manicule** (recommended, picked): the pointing hand readers drew in margins is the mark, Literata carries both
  44. display and text, and the palette is press black with one rubric red. **B · Full Stop**: a quiet lowercase name that
  45. ends on a blue proofreader's full stop. **C · Close**: an Old Town passage under a tenement, on a dark ground.
  46. | Kit: cover | Kit: typography |
  47. |---|---|
  48. | ![Kit cover](docs/images/fernhill-kit-cover.jpg) | ![Typography page](docs/images/fernhill-kit-type.jpg) |
  49. ### Tessellate: an open-source data pipeline orchestrator
  50. *Brief: precise, calm, candid; the product shows run times and costs, so figures must be tabular. Competitors: an
  51. established scheduler with a pinwheel logo, two start-ups in purple and teal.*
  52. ![Three directions for Tessellate with the comparison table](docs/images/tessellate-board.jpg)
  53. **A · Truchet Tile** (recommended, picked): one tile cut by two quarter arcs that join any neighbour however it is
  54. turned, the way steps of a pipeline meet. **B · Readable Run**: a text serif wordmark whose first *t* carries a
  55. duration bar. **C · Night Floor**: two L-pieces lock into one square, dark-first.
  56. | Kit: logo on colour, and misuse specific to this mark | Kit: applications |
  57. |---|---|
  58. | ![Logo on four grounds with contrast ratios, and six misuses](docs/images/tessellate-kit-logo.jpg) | ![Site header, business card, avatar and app icon](docs/images/tessellate-kit-applications.jpg) |
  59. ### More runs
  60. | Brief | Languages | Directions (recommended first) |
  61. |---|---|---|
  62. | ![](docs/images/gallery/molino.jpg) **Molino Coffee**, a specialty coffee shop in Lisbon | en, pt | Vizinho · Mó · Bloco |
  63. | ![](docs/images/gallery/hartwell.jpg) **Hartwell & Oduya**, a commercial law firm in London | en | The Bar Between · Equal Measure · Counterparts |
  64. | ![](docs/images/gallery/parallax.jpg) **Parallax Labs**, AI planning agents for supply chains | en | Depth Field · Parallax Error · Computation Pad |
  65. | ![](docs/images/gallery/fauvel.jpg) **Domaine Fauvel**, a natural wine estate in Burgundy | en, fr | La Taille · Climat · Fauve |
  66. | ![](docs/images/gallery/moonvault.jpg) **Moonvault**, a self-custody crypto wallet | en | Held Moon · Two O's · Lamp Arcade |
  67. | ![](docs/images/gallery/hasat.jpg) **Hasat**, a Turkish breakfast restaurant in Istanbul | tr, en | Tepeleme · Demli · Uzun Sofra |
  68. ## Install
  69. **Claude Code plugin** (recommended):
  70. ```
  71. /plugin marketplace add fatihaydost/brand-identity-skill
  72. /plugin install brand-identity@brand-identity-skill
  73. ```
  74. **Or as a plain skill**, from the `brand-identity.zip` of the
  75. [latest release](https://github.com/fatihaydost/brand-identity-skill/releases/latest) (unzip it into
  76. `~/.claude/skills/`), or from a clone:
  77. ```bash
  78. git clone https://github.com/fatihaydost/brand-identity-skill
  79. mkdir -p ~/.claude/skills
  80. ln -s "$PWD/brand-identity-skill/skills/brand-identity" ~/.claude/skills/brand-identity
  81. ```
  82. **Requirements:** Python 3.10+, a Chromium-based browser (Chrome, Chromium, Edge or Brave) and, ideally,
  83. [uv](https://docs.astral.sh/uv/). With uv the scripts install their own Python packages (fonttools, uharfbuzz,
  84. skia-pathops) into a cached environment on first use; nothing to `pip install`. Run
  85. `python3 skills/brand-identity/scripts/brand.py check` to see what is missing. Tested end to end on Linux, including
  86. a clean container; unit tests run in CI on Linux, macOS and Windows.
  87. **Other hosts.** The skill is a standard `SKILL.md` folder, so agents that read Agent Skills can load
  88. `skills/brand-identity`; it is tested end to end in Claude Code. In claude.ai chat the code sandbox does not allow
  89. Google Fonts by default, so type measurement and rendering fail there; use Claude Code.
  90. ## Use
  91. Ask Claude Code in plain words:
  92. - "Logo, fonts and colours for a specialty coffee roaster in Lisbon."
  93. - "New type and palette for https://example.com, keep the logo."
  94. - "Just a type system for my newsletter."
  95. - "Critique our brand identity."
  96. It asks one round of questions (what to make, how many sets, languages, competitors), builds the sets, shows the
  97. board and **stops** for you to pick. It never picks for you. Then it builds the kit for your choice, or for a mix
  98. ("logo from A, palette from B").
  99. A full run to the board takes about 40 agent turns and roughly 0.9 M billed-input-equivalent tokens (measured on six
  100. briefs; see [testing](docs/testing.md#budget)).
  101. ## Why the sets don't look like every other AI brand
  102. Asked for a brand identity, plain models converge: cream ground, terracotta, a gold accent, Fraunces headings, Inter
  103. body text, a logo that hides a letter in a circle. The skill knows these defaults and asks for a reason from the
  104. brief whenever a set uses one; at least one set per run uses none. On the same six briefs:
  105. | | AI-default hits (type and palette) | sets with none |
  106. |---|---|---|
  107. | Plain model, 24 answers | 4.38 per answer | 0 of 24 |
  108. | brand-identity, 18 sets | 0.28 per set | 13 of 18 |
  109. Method and data: [`docs/testing.md`](docs/testing.md#ai-defaults).
  110. ## How it works
  111. ```mermaid
  112. flowchart LR
  113. Q[One round of questions] --> S[sets.json<br/>mechanism, mark, type, palette seeds]
  114. S --> L[logos<br/>contact sheet at 16 and 32 px]
  115. L -->|fix one defect per mark, twice| S
  116. S --> B[build<br/>palette roles, audits, cards, board]
  117. B -->|a gate failed| S
  118. B --> P{You pick}
  119. P --> K[kit<br/>PDF, logo files, tokens, font sheet]
  120. ```
  121. The agent designs; the scripts measure. Gates (contrast, licence, glyph coverage, file validity, renders) stop a set
  122. until it is fixed; warnings are heuristics and say so. Architecture: [`docs/architecture.md`](docs/architecture.md).
  123. Tests (unit, smoke, rendered-output probes, end-to-end runs): [`docs/testing.md`](docs/testing.md).
  124. ## Network and privacy
  125. The skill runs locally. It contacts:
  126. - **Google Fonts** (`fonts.googleapis.com`, `fonts.gstatic.com`) and the
  127. [google/fonts](https://github.com/google/fonts) repository on `raw.githubusercontent.com`, to download fonts and
  128. their metadata and licences for measurement and rendering;
  129. - **PyPI**, through uv, once, for its Python packages;
  130. - **your site and the competitor sites you name**, opened in a headless browser, only when you give their URLs.
  131. No telemetry; nothing is uploaded. The brand's fonts are never subset or redistributed: the kit links to them and
  132. gives embed code. Licences are checked per family (OFL by default; Fontshare and commercial faces get a licence note).
  133. **Your brief and Claude.** The skill sends nothing anywhere, but what you tell Claude (the brief, the strategy, client
  134. and competitor names) is processed by Anthropic under your plan's terms. On Free, Pro and Max, including Claude Code,
  135. retention and training follow your privacy setting
  136. ([consumer terms](https://www.anthropic.com/news/updates-to-our-consumer-terms)); under the Commercial Terms (Team,
  137. Enterprise, API) Anthropic may not train models on customer content
  138. ([Commercial Terms](https://www.anthropic.com/legal/commercial-terms)). Check your plan before pasting confidential
  139. client material.
  140. **Trademark clearance is yours.** The skill does not search trademark registers, and Anthropic's IP indemnity under
  141. the Commercial Terms excludes trademark claims based on using an output in trade or commerce. Before a logo goes into
  142. use, have it cleared in the countries where the brand will trade.
  143. ## Develop
  144. ```bash
  145. pip install -r requirements-dev.txt # requirements.txt + gflanguages (checks the bundled language data)
  146. python3 -m unittest discover -s tests
  147. python3 tools/smoke_test.py
  148. python3 tools/package_skill.py # dist/brand-identity.zip
  149. ```
  150. ```
  151. skills/brand-identity/ the skill: SKILL.md, references/, scripts/, templates/, assets/
  152. tools/ smoke test, packager, run measurement, AI-defaults experiment, catalogue builders
  153. tests/ unit tests
  154. evals/ behaviour evals and fixtures
  155. docs/ architecture, testing, images
  156. ```
  157. ## Translations
  158. [한국어 (Korean)](translations/ko/SKILL.md) — the English `skills/brand-identity/SKILL.md` is canonical.
  159. ## Licence
  160. MIT, see [LICENSE](LICENSE). Adapted code and bundled data: [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md). Brand
  161. names and measured site colours belong to their owners: [TRADEMARKS.md](TRADEMARKS.md). The example briefs in this
  162. README were invented for it; any likeness to a real business is coincidental.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.