mxa-audit-discoverability
Assesses whether search engines and AI assistants can find, read and cite a site, on a sample of its pages. It establishes first whether a crawler can read the page at all.
- Checks first, through the Playwright sweep, whether AI crawlers, which do not run JavaScript, can read the page, and reports a client-rendered page as a ceiling rather than a bad grade.
- Scores crawlability and metadata with Lighthouse, runs the site's Playwright sweep for the sitemap's entries, a real 404 and whether the page keeps the head its server sent, validates structured data with Google's tools, and reads robots.txt and citability by eye rather than scoring them.
- Reads three pages per template rather than the whole site, which is how it catches a title or preview shared across every route.
- Refuses to sell llms.txt as a visibility lever, because no major engine consumes it.
- Says what it cannot see and what a full audit would add to close each gap: the engines' own data, a whole-site crawl, bot logs and a checked set of AI answers, with an access checklist for the client and the method for each.
- Names the two levers it does not touch: what each page is for, and who links to or mentions the site.
What it uses
Default
Lighthouse
Alternatives
AhrefsSemrush
Also uses
PlaywrightSearch ConsoleBing Webmaster Tools
The two webmaster tools are read by a person, not driven by the skill.
The checks module maps Lighthouse's SEO category onto crawlability and metadata; the request-level facts come from the site's Playwright sweep and the rest is read by eye.
Install it
Already on studio Claude accounts, in the Made by Agents plugin. Anyone else:
claude plugin marketplace add madebymany/madebyagentsclaude plugin install made-by-agents@madebymanyFor Cursor, Codex and other agents, just this skill:
npx skills add madebymany/madebyagents --skill mxa-audit-discoverability