Design Production-Ready Logos With a 1,400+ Logo Reference Library

Ask for a logo and most agents hand you a generic SVG in seconds. This 3k-star skill runs the process a studio would: research, concepts, tests, a hard stop for your approval — and only then the expensive part.

Why This Skill Matters

The README's workflow runs brief → research → concepts → test & refine → checkpoint. The skill maps your words, studies category conventions in its reference library, drafts 8–12 concept one-liners, builds 3 in SVG, then audits them — 16 px legibility, one-colour versions, a competitor shelf test. It always stops at the checkpoint: one overview image with a recommendation. The full kit (colour, lockups, presentation board, icons, guidelines) is only produced after you pick a direction. The library behind it holds 1,400+ real-world SVG logos across roughly 1,200 brands, each classified by mark type, technique, geometry, subject, typography, mood, and industry — used to study construction and avoid look-alikes, never to copy, per the README.

Installation

In Claude Code (the README's recommended route):

/plugin marketplace add kaankiziltug/logo-design-skill
/plugin install logo-design@logo-design-skill

Manual alternative: clone the repo and copy skills/logo-design into ~/.claude/skills/. For Claude.ai or Claude Desktop, download logo-design.zip from the Releases page and upload it under Settings → Capabilities → Skills (logo-design-lite.zip if you hit a size limit). Gemini CLI and Codex CLI work too — copy the folder to ~/.gemini/skills/logo-design or ~/.codex/skills/logo-design.

Real Workflow: A Logo for a Savings App

1. Start with the README's own example prompt:

Design a logo for Harbor, a savings app for first-time savers. It should feel calm and safe.

2. The skill asks brief questions or states its assumptions, then searches the library for category conventions.
3. At the checkpoint you get a concept overview sheet — three directions at true 64/32/16 px sizes with a recommendation. In four of the latest ten example runs the client picked a different concept than the one recommended, which is exactly what the stop is for.
4. After your pick, the kit: svg_audit.py scores production readiness (structure, colours, complexity, near-miss angles, tiny details, centring), preview_sheet.py builds the test sheet, presentation_board.py lays the mark on six industry-specific mockups, and export_variants.py produces the favicon, app-icon, and web-manifest set.

The tooling is Python 3.8+ with the standard library only. PNG/ICO export uses whatever renderer is available — cairosvg, rsvg-convert, Inkscape, a Chromium-based browser, or macOS Quick Look.

Tips

• The skill works best with a model that can view images — it renders its own drafts to PNG and checks them before showing you anything, per the README.
• It triggers on: logo, wordmark, monogram, brand mark, app icon, favicon, rebrand, and critique requests — critiquing an existing logo.svg is a supported flow.
• You can drive the tools directly, e.g. python3 scripts/search_library.py --technique negative-space --exemplary to study construction patterns.
• The library leans simple: a median of 2 colours per logo, 75% using 3 or fewer — a useful prior when a concept starts accumulating gradients.
• 141 library entries are flagged as exemplary teaching examples, searchable with --exemplary.

When Not to Use This

The library's SVG files are trademarks of their respective owners — included for reference and education only, and explicitly not covered by the skill's MIT license, so nothing from it belongs in client work. And the scope is the mark: references extend to identity systems, but if you need a full brand rollout beyond the logo, plan for tools beyond this skill.

See the leaderboard for more skills.