Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add S3YED/appie-kit --skill building-ai-websites-that-dont-look-aigit clone --depth 1 https://github.com/S3YED/appie-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/s3yed/appie-kit/building-ai-websites-that-dont-look-ai)<a href="https://agentmods.dev/skills/s3yed/appie-kit/building-ai-websites-that-dont-look-ai"><img src="https://agentmods.dev/badge/skills/s3yed/appie-kit/building-ai-websites-that-dont-look-ai/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/s3yed/appie-kit/building-ai-websites-that-dont-look-ai"><img src="https://agentmods.dev/badge/skills/s3yed/appie-kit/building-ai-websites-that-dont-look-ai.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00062 | $0.01231 |
| Opus 5 | $0.00031 | $0.00616 |
| Sonnet 5 | $0.00012 | $0.00246 |
| Haiku 4.5 | $0.00006 | $0.00123 |
Grade A, and why
building-ai-websites-that-dont-look-ai scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Building AI Websites That Don't Look AI-Generated
Distilled from a 67-min masterclass (YT 2Gda_ZvV1V4, transcript at ~/clawd/knowledge/webdesign-video-2Gda_ZvV1V4/transcript.txt). Core thesis: AI sites look like slop because people start from a blank canvas. Fix: extract a real brand's design system first, merge it with a proven layout structure, then layer design fundamentals on top.
The key insight
Separate brand from structure — they come from different sources:
- Brand kit (colors/type/radii/shadows): extract from a reference site's screenshots or a pre-built kit (Get Design has 71 brand kits as
design.md) - Page structure: a Dribbble screenshot that fits the BUSINESS TYPE (don't copy Stripe's software-company layout for a service business)
- Merge both in Claude Design: brand system + structure screenshot → hi-fi multi-page prototype → hand off to Claude Code ("copy it pixel for pixel", Next.js + Tailwind)
Workflow (17 steps, condensed)
- Pick style reference (screenshots) + structure reference (Dribbble: search "[niche] website")
- Claude Design > Design Systems > upload screenshots or design.md > publish
- Generate hi-fi prototype: name the pages, "use the brand design system for branding, use the attached screenshot for page structure"
- Hand off to Claude Code; CLAUDE.md persona: "You are a senior UI designer and front end developer. You build premium, modern, elegant interfaces."
- Layout: 6 decisions = direction, column ratio (60/40, 70/30 beats equal), container width, section rhythm, padding, alignment. Generous padding, applied consistently to ALL sections
- Typography: 3 distinct font families (heading dark/bold/large, body lighter, caption in accent color); gradient on key heading words; roomy line-height
- Imagery: Pixabay (royalty-free) + Freepik illustrations + Lucide icons. Never leave placeholders
- "Expensive" backgrounds = LAYERS: solid + radial gradient + subtle noise/dot texture + grid + color blobs + frosted glass. "People won't see the noise, but it fills out the page"
- Shadows = elevation (3D leap off the page); border radius identical everywhere (from the kit)
- Hover/focus states with 300ms transitions (never instant): buttons scale +10% + shadow, nav links get left-to-right underline, inputs get ring + ring offset, cards lift
- Nav: sticky + frosted glass; mobile hamburger with slide-out
- Dark/light toggle (optionally default by user timezone)
- Motion: scroll-triggered pop-ins with STAGGERED delays (never simultaneous), logo marquee, count-up numbers, shimmer, floating elements. Prompt lever: "10x the amount of animations on this page", then dial back to taste
- 3D hero: remix a Spline community scene, transparent bg, brand colors, export embed (caveat: ~5MB, hurts page speed)
- AI video: Higgs Field MCP (
http://mcp.higgsfield.ai, Seedance 2.0) — video bg also costs load speed - Responsive: "optimize for desktop, tablet and mobile" + verify in DevTools device emulation. Mobile first: 70% of traffic
- Deploy: push to GitHub (private) > Vercel import > custom domain
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 63 lines · 62 tokens per session scan A 37ef015a27b9
building-ai-websites-that-dont-look-ai is a skill published in the GitHub repository S3YED/appie-kit (9 stars, last pushed 17d ago), licensed MIT. It adds 62 tokens to every session and 1,231 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
animation-patterns
Framer Motion patterns, page transitions, skeleton loading, scroll-linked animations, and gesture-based interactions for React.
tailwind-theme-builder
Set up Tailwind v4 + shadcn/ui themed UI with dark mode. Install deps, configure CSS variables via @theme inline, wire dark mode toggle, verify. Use whenever the user mentions Tailwind v4, setting up Tailwind theming, shadcn/ui colours, dark mode, or troubleshooting colours not working, tw-animate-css errors, @theme…
generic-react-design-system
Complete design system reference for React applications. Covers colors, typography, spacing, component patterns, glassmorphism effects, GPU-accelerated animations, and WCAG AA accessibility. Use when implementing UI, choosing colors, applying spacing, creating components, or ensuring brand consistency.
Premium Web Design
Comprehensive luxury web design skill with reusable animation/interaction patterns, reference site analyses, and tool guides for building stunning React+TS+Tailwind websites.
21st.dev Magic MCP
Generate premium React+TS+Tailwind UI components from natural language via 21st.dev Magic MCP server.
ui-design
Designs and builds React/Next/Tailwind UI and audits visual and interaction defects. Use when asked to "build a landing page", "extract our design system", "add dark mode", "make this responsive", "remove UI slop", or "audit this component". For product decisions use product-design; for browser measurements use…