juliobsz/skills

An agent skill that builds award-caliber websites from one line

What it is

/awwwards is a skill (plugin) for AI‑agent toolchains such as Claude Code, Codex, Cursor, Gemini‑CLI, Copilot, etc. When you give an agent a one‑sentence brief like “make me a portfolio, my name is Julio” the skill drives the agent through a full web‑site creation pipeline – from concept generation, design‑system definition, code generation, automated QA, to optional deployment on Vercel/Netlify/Cloudflare.

How it works (high‑level)

  1. Brief parsing – the model extracts the client’s name, craft, location and any project details, filling missing facts with written assumptions.
  2. Concept generation – three distinct concepts are produced, scored for originality/fitting the brief; the best becomes the site’s core idea and signature animation.
  3. Art‑direction – a palette, free‑font pairings, grid, rhythm and motion language are written to a DESIGN.md. A “genericness test” discards bland AI‑default designs.
  4. Build – semantic HTML (Next.js by default, or Astro/any existing stack) is generated, with motion powered by GSAP, Lenis (smooth scroll) and OGL (WebGL). Motion is optional and respects reduced‑motion settings.
  5. QA loop – the generated site is built, then a sub‑agent runs a crawling audit (SEO, accessibility via axe‑core, weight, placeholder detection) and takes real‑browser screenshots (desktop, mobile, reduced‑motion). It scores the site against a rubric (aim ≥ 8/10) and iteratively fixes issues.
  6. Ship – creates OG image, favicon, sitemap, robots.txt and, after user confirmation, deploys to a cloud host.

What you get

  • A complete, production‑ready website tailored to the brief.
  • Design assets (palette, typography, motion specs) documented in DESIGN.md.
  • Automated quality checks (accessibility, performance, SEO) with a clear score.
  • Optional deployment scripts for Vercel, Netlify or Cloudflare.
  • Stand‑alone helper scripts (audit.mjs, shoot.mjs, og.mjs, etc.) that can be run on any site.

Who it’s for

  • Designers or developers who want an AI‑driven “creative director” to produce a high‑quality portfolio, product landing page, or studio site.
  • Teams using Claude Code, Codex, Cursor, Gemini‑CLI, Copilot, or any agent that supports the skills CLI.
  • Users comfortable running Node/Playwright scripts locally and who can review placeholder content before publishing.

Limitations

  • The process is slow and token‑heavy; it’s meant for a full session, not a quick snippet.
  • Quality depends on the underlying LLM (tested with Claude Code and Codex).
  • Generated copy may contain placeholders that need manual replacement.
  • Requires Node ≥ 22.20, Playwright, axe‑core, and a model capable of image‑based reasoning.

Getting started

# install the skill globally (or into a specific agent)
npx skills add juliobsz/skills -g
# start a new agent session and invoke the skill
/awwwards   # or just ask for a website

You can also install the optional /anatomy skill for interactive isometric explainer figures.

All information above is taken directly from the repository’s README; no additional features are inferred.

Related

  • Project
    galilai-group/stable-pretraining

    stable‑pretraining is a PyTorch‑Lightning (and experimental JAX) library that lets you plug‑in self‑supervised learning methods by defining only a forward function. It handles data loading (with optional GPU‑side augmentations), checkpointing, multi‑GPU training, and a rich set of online evaluation callbacks, making SSL research faster and more reproducible.

  • Project
    evotai/evot

    Evot is a lightweight terminal‑based coding agent that lets a LLM read, edit, write files and run shell commands. With a tiny prompt, four core tools, background execution, and plug‑in support for many LLM providers, it offers fast, low‑cost AI‑assisted programming directly from the command line.

  • Project
    PinelliaChill/Operant

    Operant is a local, extensible AI‑agent platform that lets a single user run multiple LLMs in coordinated workflows (planning, coding, reviewing, etc.). It offers safety‑first approvals, a versioned memory system for re‑using experience, plugin‑based tool extensions, and several interfaces (CLI, web UI, TUI, desktop). All state is stored locally and model calls go to any OpenAI‑compatible API you configure. The project is in beta (macOS arm64 & Ubuntu 24.04 validated) and is licensed for non‑commercial use only.

  • Project
    LucieEveille/kiwi-mem

    kiwi‑mem is a self‑hosted FastAPI gateway that adds human‑like long‑term memory to LLM chats. It stores facts in PostgreSQL + pgvector, assigns each a heat that decays over time, merges related pieces during a nightly “Dream” cycle, and compresses old chats into hierarchical calendar summaries. The system works with any OpenAI‑compatible or Anthropic native API, provides a web admin panel, and runs via a single Docker‑Compose command.