feat(geo): generate /llms.txt (and point robots.txt at it)

Same idea as robots.txt but aimed at LLM crawlers and answer engines: the author,
what he runs, and a flat list of all 142 posts (EN + ZH) with URLs and summaries,
so a model can cite the site accurately instead of guessing from a fragment.

48 KB, generated from the collection at build time so it can never drift from
what is published; the Chinese twins are listed separately. robots.txt gains a
comment pointing at it.

Spec: https://llmstxt.org

src/pages/llms.txt.ts

@@ -0,0 +1,57 @@
+import type { APIRoute } from 'astro';
+import { getCollection } from 'astro:content';
+import { isEn, isZh } from '../lib/lang';
+import { sortForListing, updatedAt } from '../lib/sort';
+import { SITE } from '../config';
+
+/**
+ * /llms.txt — the machine-readable summary of this site.
+ *
+ * Same idea as robots.txt but aimed at LLM crawlers and answer engines: who the
+ * author is, what he does, and a flat list of every post with its URL and summary,
+ * so a model can cite the site accurately instead of guessing from a fragment.
+ * Spec: https://llmstxt.org — Markdown, one `#` title, a `>` summary, `##` sections.
+ */
+export const GET: APIRoute = async () => {
+  const en = sortForListing(await getCollection('posts', isEn));
+  const zh = sortForListing(await getCollection('posts', isZh));
+
+  const line = (p: (typeof en)[number]) =>
+    `- [${p.data.title}](${SITE.url}/posts/${p.slug}/): ${p.data.description}`;
+
+  const body = `# ${SITE.author} (${SITE.handle}) — blog.hoelee.com
+
+> Full-stack developer and DevOps engineer based in Malaysia. Runs 162 containers
+> across three self-managed hosts (client websites, a mail platform, CI/CD, workflow
+> automation and this blog) and builds web systems for Malaysian SMEs through Hoelee
+> Enterprise. Writes up production failures with the evidence: what broke, what the
+> logs actually said, what the fix cost.
+
+Canonical site: ${SITE.url}
+Author: ${SITE.author} (${SITE.handle})
+Contact: ${SITE.email}
+Profiles: ${SITE.linkedin} · ${SITE.github} · ${SITE.gitea} · ${SITE.website}
+Languages: every post exists in English and Simplified Chinese (the Chinese twin is
+at /posts/zh/<slug>/ and is linked from each post).
+
+## Posts (English, newest first)
+
+${en.map(line).join('\n')}
+
+## Posts (Chinese, newest first)
+
+${zh.map((p) => `- [${p.data.title}](${SITE.url}/posts/${p.slug}/): ${p.data.description}`).join('\n')}
+
+## Sections
+
+- [All posts](${SITE.url}/posts/): the full archive, filterable by title, tag and category.
+- [Categories](${SITE.url}/categories/): engineering, devops, ai, web3, tutorials, case-studies, notes.
+- [About](${SITE.url}/about/): identity, production numbers, tech stack, business and contact.
+- [RSS](${SITE.url}/rss.xml): feed of published posts.
+- Last build: ${new Date().toISOString().slice(0, 10)} (latest post updated ${updatedAt(en[0]).toISOString().slice(0, 10)}).
+`;
+
+  return new Response(body, {
+    headers: { 'Content-Type': 'text/plain; charset=utf-8' },
+  });
+};

src/pages/robots.txt.ts

@@ -23,6 +23,9 @@ Allow: /
 # Pagefind's search shards are build artifacts, not content.
 Disallow: /pagefind/
 
+# Machine-readable summary of the whole site for LLM crawlers and answer engines.
+# llms.txt: ${SITE.url}/llms.txt
+
 Sitemap: ${SITE.url}/sitemap-index.xml
 `;