<?xml version="1.0" encoding="UTF-8"?>
<!--
  Written 2026-09-05, when the origin opened to indexing.

  It did not exist before then, deliberately: robots.txt served Disallow: /
  and a sitemap listing pages a crawler is forbidden to fetch is noise. The
  old robots.txt carried a commented Sitemap line and a note saying to write
  the file before uncommenting it. This is that file.

  LASTMOD IS A CLAIM, so it is only moved when the page really changed. It is a
  hint to Google that a re-crawl is worth its time, and a date that is not true
  is a hint it learns to ignore. That cuts both ways: a date left BEHIND is a
  false claim too. All three pages sat at 2026-09-10 until 2026-09-15, although
  the copy rewrite and the 31-item UI pass changed every one of them on
  2026-09-12 - and Google's last crawl of the homepage was 2026-09-10, so the
  file was telling it there was nothing new to fetch.
  Current dates: /, /r6-recoil and /aim-assist 2026-09-26, when every footer
  gained a link to /download (before that, nothing on the site linked to it).
  Earlier moves: 2026-09-15 for the owner's titles and descriptions, and for
  /aim-assist the pre-render and the WebGL gate that let Google render it.
  BUMP THE DATE IN THE SAME CHANGE THAT CHANGES A PAGE's visible content.

  IT IS HAND-MAINTAINED. Nothing derives it from git, so it goes stale silently
  - which is fine for three documents and would not be for thirty.

  FOUR URLS, because there are four documents. /checkout and /admin are
  excluded here and in robots.txt - one is a transaction, the other is a
  console. /aim-assist is a separate React application served from the same
  origin, which is exactly why it belongs in this file: one origin, one
  sitemap, whatever the build looks like underneath.

  /download JOINED ON 2026-09-12, when it stopped being a static file in
  public/ and became a real page. robots.txt has always allowed it; it was
  simply never listed, so the one page a buyer is sent to after paying was the
  one page Google had no pointer to. Its lastmod is the day it was rebuilt.
-->
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>https://miyu.lol/</loc>
    <lastmod>2026-09-26</lastmod>
    <changefreq>monthly</changefreq>
    <priority>1.0</priority>
  </url>
  <url>
    <loc>https://miyu.lol/r6-recoil</loc>
    <lastmod>2026-09-26</lastmod>
    <changefreq>weekly</changefreq>
    <priority>0.9</priority>
  </url>
  <url>
    <!-- NO TRAILING SLASH. The Worker serves assets with
         html_handling: "drop-trailing-slash", so /aim-assist/ answers 307 and
         redirects here. A sitemap entry or a canonical pointing at a redirect
         spends crawl budget to arrive at a URL it was not told about, and on a
         site with three pages that is the whole budget. Measured 2026-09-07:
         /aim-assist 200, /aim-assist/ 307. -->
    <loc>https://miyu.lol/aim-assist</loc>
    <lastmod>2026-09-26</lastmod>
    <changefreq>weekly</changefreq>
    <priority>0.9</priority>
  </url>
  <url>
    <!-- Lower priority than the two product pages on purpose: this is where
         somebody goes AFTER deciding, so it should never outrank the pages
         that do the deciding. Weekly, because a release changes what it
         serves without changing a word on it. -->
    <loc>https://miyu.lol/download</loc>
    <lastmod>2026-09-12</lastmod>
    <changefreq>weekly</changefreq>
    <priority>0.6</priority>
  </url>
</urlset>
