A deep skill that runs a full logo design process (brief, concepts, SVG construction, testing, kit) backed by a 1,400+ logo reference library and Python tools.
A deep, well-tested single skill with a large, properly licensed reference library and a sensible checkpoint before heavy output; actively maintained.
This is one skill, logo-design, but it is large: about 17.6 KB of instructions, 14 reference files and nine Python scripts, built around a reference library of over 1,400 real SVG logos classified by mark type, technique, geometry, subject, typography, mood and industry. The workflow runs discovery, word mapping, choosing a mark type (wordmark, monogram, letterform, pictorial, abstract, emblem, mascot, combination), concepting, geometric SVG construction, optical corrections, color and typography, then testing.
The skill always stops at a checkpoint after building and testing 3 concepts, shown as one overview image with a recommendation; nothing further (color systems, lockups, presentation boards, icon exports) happens until you pick a direction. Its Python tools (standard library only, PNG export falling back across available renderers) audit an SVG against logo principles, build concept overview sheets and test sheets (16px pixel test, one-color, reversed, mirror, shelf test against competitors), generate client presentation boards with industry mockups, and export a full favicon and web-manifest set. The library is explicitly for studying construction and avoiding look-alikes, not for copying. It also has a critique mode, a redesign mode and an identity-system mode for guidelines and sub-brands. The repo has CI tests and a lite zip without the SVG library for size-limited uploads.
One skill with 14 reference files, 9 scripts, a 1,400+ file SVG library, and CI tests — references, scripts and tests together, the top of the depth scale.
Counted from the repository file tree. How we score depth.
A logo-design skill with reference guides and nine local Python scripts that build, audit and render SVG logos and preview sheets. The one script that starts other programs launches a local image renderer or a headless browser on a temporary profile; none of the scripts we looked at use the network or credentials. SKILL.md sets no allowed-tools. Only render_png.py was read in full; the large icon library was not available.
The repository (1560 files) on Oct 29, 2026. Not a full review: 1445 of 1560 files were not downloaded (large repository; we reviewed the skill folders); we read 1 of 9 code files; it relies on code outside the repository: 1445 of 1560 repository files were not available to us (the bundled SVG icon library, plugin manifests, workflows); optional renderers cairosvg, rsvg-convert, inkscape or a browser are used if present.
NVIDIA SkillSpector static scan: 13 pattern hits (Dangerous Code Execution 7, Anti-Refusal 2, Data Flow 2, Privilege Escalation 1, Excessive Agency 1). All 13 SkillSpector hits are in render_png.py: subprocess calls to image renderers and a browser with a temporary profile (the 'keychain' hit is Chrome's --use-mock-keychain flag), a taskkill of that browser process, plus 'never judge' wording in design guidance.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT for skill text, scripts, templates and catalog data; the bundled SVG logo files in assets/library/svg/ carry their own license noted in TRADEMARKS.md, and the example brands are fictional.
Not really. It deliberately pauses at a checkpoint for human sign-off before producing the full kit; it is not meant to run unattended.
Points a headless browser at any site and reconstructs its design system as 8+ files: DTCG tokens, Tailwind config, shadcn theme, Figma variables, WCAG score.
A design skill with 24 commands (polish, bolder, critique, audit, harden and more), a 62-rule deterministic slop detector and an optional edit hook.
Nine skills for design engineers: a UI baseline checklist, accessibility/metadata/motion fixers, a design.md generator, and canvas builders for components.
It is a single comprehensive Claude/Codex/Gemini CLI skill that runs a full logo design process from brief to production SVG files, backed by a library of over 1,400 classified real-world logos and Python tools that audit and test designs.
In Claude Code, run /plugin marketplace add kaankiziltug/logo-design-skill then /plugin install logo-design@logo-design-skill, or clone the repo and copy skills/logo-design into your agent's skills directory. A size-limited lite zip without the SVG library is also provided.
No. The README says the reference library is used to study construction and category conventions and to avoid look-alikes, never to copy. It builds new SVG geometry for each concept.
The tools are standard-library Python with no network calls described, and the repo has CI tests. The bundled logo library has its own license terms separate from the MIT-licensed skill code.