Use getSiteUrl() instead of re-implementing the site URL fallback in sitemap, robots, and feed

[royalpinto007]

Why it matters

lib/utils/urls.ts already exports getSiteUrl(), which reads NEXT_PUBLIC_SITE_URL, falls back to NEXT_PUBLIC_APP_URL, strips trailing slashes, and only then falls back to the production domain. But several files re-implement a weaker version of that logic inline:

const siteUrl = process.env.NEXT_PUBLIC_SITE_URL ?? "https://agentpostmortem.com";

That copy ignores NEXT_PUBLIC_APP_URL and does not strip a trailing slash, so a preview deploy configured with https://preview.example.com/ produces double-slash URLs in the sitemap and RSS feed.

Where

Replace the inline fallback with import { getSiteUrl } from "@/lib/utils/urls" in:

• app/sitemap.ts (line ~10)
• app/robots.ts (line ~5)
• app/feed.xml/route.ts (line ~8)
• app/(public)/agent/[slug]/page.tsx (line ~21)
• app/(public)/tag/[slug]/page.tsx (line ~21)
• app/(public)/case/[caseNumber]/page.tsx (lines ~44 and ~135)

app/layout.tsx and lib/resend/send.ts already do it the right way, so use those as the reference.

Notes

• Leave the hardcoded agentpostmortem.com strings that are display text or mailto: addresses alone. This is only about URL construction.
• Run npm test and npm run format before opening the PR.

Scope is a handful of one-line edits, comfortably under an hour.

Questions are very welcome. Comment here to claim it and ask anything you are unsure about, you will usually get a reply within a day.

[royalpinto007]

Thanks @slegarraga, #22 is up and in my review queue.