{
  "version": "1.0",
  "updated": "2026-05-06",
  "skills": [
    {
      "name": "a14y",
      "description": "Use when the user wants to audit a website for agent readability, see how well AI agents can discover/parse/comprehend a site, run the a14y CLI, evaluate llms.txt/AGENTS.md/sitemap quality, or improve a project's a14y score. Detects a running local dev server first, falls back to the live URL, runs the audit with `--output agent-prompt`, then presents an itemized plan the user accepts or skips per fix before any change is made, and tracks score history in AGENTS.md.",
      "url": "/.well-known/agent-skills/a14y/SKILL.md",
      "license": "Apache-2.0",
      "compatibility": "Requires Node.js 18+ for `npx`, network access to reach the target URL, and shell access. Works with any coding agent that supports the agentskills.io spec (Claude Code, Codex, Cursor, OpenCode, etc.).",
      "metadata": {
        "version": "0.2.0",
        "homepage": "https://a14y.dev",
        "source": "https://github.com/timothyjordan/a14y",
        "spec": "https://agentskills.io/specification"
      }
    }
  ]
}