chore(web): keep settings out of search results

/settings is a per-viewer preferences page with nothing to index, and it
sets robots noindex already; robots.txt now says so too, next to /me.

Also records why character pages are not in the sitemap: like sticker
pages they are linked from every pack page, and the distinct characters
can only be found by walking every pack's detail.

Claude-Session: https://claude.ai/code/session_01WYeKk1xWGbYeENdb6Ez3xL

Author: KunMoe

apps/web/public/robots.txt

@@ -6,6 +6,9 @@ Disallow: /ja/auth
 Disallow: /me
 Disallow: /en/me
 Disallow: /ja/me
+Disallow: /settings
+Disallow: /en/settings
+Disallow: /ja/settings
 Disallow: /api
 
 Sitemap: https://sticker.kungal.com/sitemap_index.xml

apps/web/server/api/__sitemap__/urls.ts

@@ -41,9 +41,12 @@ export default defineCachedEventHandler(
     const urls: SitemapEntry[] = [entry('/', 1), entry('/about', 0.6)]
     const apiBase = useRuntimeConfig().apiBaseUrl
 
-    // Individual sticker pages are deliberately absent: they are linked from
-    // every pack page, so crawlers still reach them, and listing them here grew
-    // the file by one URL per sticker per locale with no ceiling.
+    // Individual sticker pages and character pages are deliberately absent.
+    // Both are linked from every pack page, so crawlers still reach them, and
+    // neither can be enumerated cheaply: stickers grow the file by one URL per
+    // sticker per locale with no ceiling, and the distinct characters are only
+    // discoverable by walking every pack's detail. Add a listing endpoint first
+    // if character pages ever need to be in here.
     try {
       const authors = new Set<number>()