Curated design-agent skills, rubrics, rankings, and examples for higher-quality AI-generated UI and brand systems.
SaferSkills independently audited awesome-design-agent-skills (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<div align="center">
<img src="./hero.jpg" width="100%" alt="Awesome Design Agent Skills Hero Banner" />
Ranked design skills for coding agents: generation, polish, brand systems, and audits.
</div>
<!-- earned-skill-index:2026-08-30 -->
Install a few named design skills. Do not bulk-install a registry.
| Pack | Job |
|---|---|
| bergside/awesome-design-skills | Ranked DESIGN.md / SKILL.md directory — pick files, don't copy all 67 |
| anthropics/skills | Official frontend-design, canvas-design, theme-factory |
| nextlevelbuilder/ui-ux-pro-max-skill | Pattern intelligence for product UI |
| pbakaus/impeccable | Critique and anti-pattern detection |
| garrytan/gstack | Design review + visual QA loops |
| NVIDIA/SkillSpector | Scan a skill before it runs |
Safety and the 5–7 doctrine: awesome-hermes-agent-skills.
Brand and design-agent stack
The 2026 comparison is not another unverified prompt list. It separates discovery, specialist acceleration, owned brand authority, rendered inspection, and production proof.
Every agentic design skill on GitHub, live-verified per harness: [registry/](./registry/README.md) — Claude Code · Codex · Gemini CLI + Antigravity · Grok · Cursor · cross-harness standards · registry.json (machine-readable).
The headline finding: the three biggest this-repo design skills (ui-ux-pro-max, taste-skill, impeccable) are harness-agnostic SKILL.md packs — quality now follows the installed pack, not the harness. Parent-pack star counts (e.g. frontend-design inside anthropics/skills) are labeled in the registry and never used for scale or rank. See star-count honesty.
https://frankxai.github.io/awesome-design-agent-skills/). Desktop and mobile plates compile from registry.json; OG image rasters from the same data.node scripts/slop-scan.mjs <file> · scripts/qa-probe.js · node scripts/eval-instruments.mjs · node docs/site/build-plate.mjs · python scripts/render-og.py| Job | Recommended leader | Use it for |
|---|---|---|
| Pattern intelligence | UI UX Pro Max | Searchable styles, product types, palettes, typography, UX rules, and stack guidance |
| Frontend critique and polish | Impeccable | Product/brand shaping, browser iteration, 23 commands, deterministic anti-pattern detection |
| Concise frontend baseline | Anthropic frontend-design | Subject-grounded composition, type, hierarchy, and signature detail |
| High-variance marketing UI | Taste Skill | Opinionated anti-slop direction for landing pages and portfolios |
| Interaction craft | Emil Kowalski skills | Motion frequency, easing, interruption, microinteractions, and reduced motion |
| DESIGN.md extraction | ibelick/ui-skills | Evidence-led design-system recovery |
| UI compliance | Vercel Agent Skills | Current web-interface guideline audit |
| Owned authority | Starlight Design Intelligence | Brand packs, logo/type rights, quality gates, approval, and release-contract requirements |
[!IMPORTANT] No public skill is an end-to-end brand authority. Generated logos are not vector masters; font presence is not licensing; screenshots are not production proof; stars are not quality scores. Use the full evidence and release loop.
<!-- Keep the two GitHub callouts distinct for markdownlint MD028. -->
[!TIP] New to agent skills? Start with the Getting Started Guide, then use the anti-slop rubric on a rendered result.
starlight-design-intelligence is the owned design authority and release-contract layer.We welcome contributions that align with our core thesis of verifiable, world-healing, and rigorously-tested agent intelligence.
Please read our CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests.
This list is dedicated to the public domain under CC0 1.0.
<div align="center"> <sub>Built with ❤️ by the FrankX Superintelligence System</sub> </div>
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.