# mrpectus — https://mrpectus.com
#
# Cloudflare was serving an auto-generated content-signals file here, which
# carried no Sitemap: line and none of this site's own rules. This replaces it.

User-agent: *
Allow: /

# Paid deliverables. These already answer 401 without an unlock cookie (see
# functions/_middleware.ts), so this is belt-and-braces — it stops the URLs
# being crawled and listed at all.
#
# The four below are listed one by one on purpose. A `-program` suffix does not
# mean a page was paid for: /programs/pe-rib-flare-program is a free article,
# and a rule written to that pattern hid it from search. The paid list lives in
# GATED (src/data/gated.ts); these lines and NOT_INDEXABLE in astro.config.mjs
# have to match it.
#
# Prefix rules, deliberately unanchored: the pages answer on both
# /programs/x-program and /programs/x-program/, and a `$` would only cover one.
Disallow: /guides/*/read
Disallow: /programs/*/program
Disallow: /programs/pectus-excavatum-program
Disallow: /programs/pectus-carinatum-program
Disallow: /programs/rib-flare-program
Disallow: /programs/confidence-program

# The password prompt itself, the component gallery, and the member app.
Disallow: /unlock
Disallow: /design-system
Disallow: /app

# The post-purchase thank-you page and the page that reopens a lead's own saved
# assessment from their email are NOT listed here on purpose, even though both
# match NOT_INDEXABLE in src/data/indexable.mjs. They now carry a `noindex` meta
# instead (QuizLayout takes the prop), and a Disallow would defeat it: a page
# Google may not crawl is a page whose noindex Google never reads, so an inbound
# link can still put a URL-only result in the index. Blocking and noindexing the
# same page is the one combination that does not work. See docs/seo.md.

# API endpoints and the video proxy — nothing to index, and crawling the video
# route would pull tens of gigabytes out of R2.
Disallow: /api/
Disallow: /video/

Sitemap: https://mrpectus.com/sitemap-index.xml