Resources

A Claude Code skill that helps developers identify and fix WCAG violations in web applications. It provides guidance on semantic HTML, ARIA attributes, keyboard navigation, and color contrast ratios for both manual reviews and automated audits.

A gallery of products that support the agent skills format.

Anthropic's implementation of skills for Claude.

Web framework for building fast, content-focused websites with a flexible component model that works with React, Vue, Svelte, or plain HTML. A solid pick for portfolios, blogs, and marketing sites where performance and SEO matter more than heavy client-side interactivity.

1000+ agent skills from official dev teams (Anthropic, Vercel, Google Labs, Figma, Stripe) and the community, compatible with Claude Code, Codex, Cursor, and Gemini CLI. Broader than the travisvn list, with a cross-tool focus.

Curated list of skills, resources, and tools for customizing Claude workflows. Good jumping-off point when you want to find something specific rather than read through repos one by one.

Low-fidelity wireframing tool that keeps designs intentionally rough so stakeholders focus on structure over aesthetics.

AI app builder from StackBlitz. Prompts generate full-stack apps that run in the browser. Similar lane to Lovable. Good for fast prototypes without a local setup.

Open-source, Claude Code-powered job search system that evaluates job fit against your CV, generates tailored ATS-optimized PDFs, and scans 45+ company career pages. Built by someone who used it to land a Head of Applied AI role after evaluating 740+ offers.

A Claude Code skill that runs comprehensive accessibility audits on web projects using axe-core for runtime DOM testing and jsx-a11y for static analysis. It checks WCAG 2.1 Level A and AA compliance, flagging common issues like missing alt text, insufficient color contrast, broken heading hierarchy, and invalid ARIA attributes.

Terminal-first AI pair programmer. My primary building environment.

A full-stack React Native framework with cloud services for building, deploying, and updating mobile apps across iOS, Android, and web from a single codebase.

Still the default for collaborative interface design. Good for visual refinement after wireframes have done their job.

Guide to Figma MCP server, which lets AI agents read design context from Figma files, write directly to the canvas, and turn live web UI into editable design layers. Works with Claude Code, Cursor, VS Code, and other MCP clients via either a remote server or the desktop app.

Google Gemini's Command Line Interface for code generation.

Converts any website into fully editable Figma designs without building each element from scratch.

The design language that makes your AI harness better at design.

A design brief and specification builder that walks you through discovery, problem and solution exploration to generate a build-ready spec, step by step.

AI app builder that turns prompts into working full-stack apps. Good for spinning up prototypes fast without writing boilerplate.

A browsable directory of Agent Skills and MCP servers that extend AI coding agents like Claude Code, Claude.ai, Codex, and ChatGPT. You can discover skills across categories like developer tools, security, productivity, testing, and accessibility, with each listing showing the skill's purpose and implementation details.

Official directory of Model Context Protocol servers. The extension layer that makes Claude Code actually useful for real work.

Motivist is a goal tracking and motivation app.

Zero-config deploys for Astro. What this site runs on.

A profile-driven Claude Code skill for job search and interview prep, built around 6 role profiles that drive discovery, fit scoring, resume generation, and post-mortem learnings from a single structured career inventory. Every interview cycle feeds back into the profile, so each round starts smarter than the last.

Open-source browser automation framework from Microsoft that drives Chromium, Firefox, and WebKit through a single API for end-to-end testing and scripted browser tasks.

Skills for a personal "second brain" that runs entirely on plain markdown files maintained by Claude Code

A pluggable linting tool that scans your project for committed credentials like API keys, private keys, and tokens before they hit your repo. Runs as a pre-commit hook or CI check, with built-in rules for AWS, GCP, GitHub, OpenAI, Anthropic, Slack, and more.

Bill Buxton on why sketches are the vehicle, not the destination. Foundational reading if you're trying to understand why low fidelity matters.

An open-source secrets scanner that finds, verifies, and analyzes leaked credentials across Git repos, cloud storage, Docker images, filesystems, and SaaS tools like Postman and Jenkins. Supports over 800 credential types with live API verification, so you know whether a leaked key is actually exploitable and not just a false positive.

Claude skill that pulls UX sentiment and friction patterns for any product and turns them into interview-ready talking points. Good for design interview prep and competitive analysis.

An agent skill that audits AI-generated code for common security vulnerabilities that coding assistants tend to introduce, like hardcoded secrets, missing row-level security, client-submitted prices, and tokens in localStorage. Built for Claude Code and other Agent Skills-compatible tools, it loads only the checks relevant to your stack (Supabase, Stripe, React Native, etc.) so you're not wasting context on irrelevant rules.

The book I co-wrote with Leon Barnard and Billy Carlson. On using wireframes as a tool for thinking and collaboration, not just UI design.