SEO — Sitemap Skill | Agent Skills

SEO — Sitemap

Overview

A Sitemap is an XML file (or RSS/Atom/text alternative) that lists URLs on your site so search engines can discover and crawl them more efficiently. It supplements link-based discovery; it does not guarantee indexing.

Principles: UTF-8 encoding; absolute URLs with protocol; one host per sitemap; accurate lastmod; stay under size limits; place at site root when possible.

Reference: sitemaps.org protocol · FAQ

When to use a sitemap

| Scenario | Approach | |----------|----------| | Marketing / docs site | Root sitemap.xml with static + dynamic URLs | | Large site (>50k URLs) | Multiple sitemaps + sitemap index | | CMS with many pages | Generate from DB/CMS at build or request time | | Recent changes only | Split volatile URLs; use lastmod in index for incremental fetch |

Pair with robots.txt (Sitemap: https://example.com/sitemap.xml) and Search Console submission.

XML format (protocol 0.9)

Required structure:

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>https://www.example.com/</loc>
    <lastmod>2025-06-13</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.8</priority>
  </url>
</urlset>

| Tag | Required | Notes | |-----|----------|-------| | <loc> | Yes | Full URL with https://; max 2,048 chars | | <lastmod> | No | W3C datetime (YYYY-MM-DD or 2004-12-23T18:00:15+00:00); page last modified, not sitemap generation time | | <changefreq> | No | Hint only: always, hourly, daily, weekly, monthly, yearly, never | | <priority> | No | 0.0–1.0 relative within your site; default 0.5; does not affect ranking vs other sites |

Entity escaping in XML: & → &, < → <, > → >, " → ", ' → '.

Limits (FAQ)

Max 50,000 URLs per sitemap file.
Max 50 MB uncompressed (52,428,800 bytes).
Use gzip compression when serving (file must still be ≤50 MB uncompressed).
Exceed limits → split into multiple sitemaps + sitemap index.

Sitemap index

<?xml version="1.0" encoding="UTF-8"?>
<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <sitemap>
    <loc>https://www.example.com/sitemap-posts.xml</loc>
    <lastmod>2025-06-13T12:00:00+00:00</lastmod>
  </sitemap>
  <sitemap>
    <loc>https://www.example.com/sitemap-pages.xml</loc>
    <lastmod>2025-06-01</lastmod>
  </sitemap>
</sitemapindex>

Index: max 50,000 sitemaps, max 50 MB. Index lastmod = when that sitemap file changed (enables incremental crawl).

File location rules

Recommended: https://example.com/sitemap.xml (root).
All <loc> URLs must share the same protocol and host as the sitemap (e.g. no http URLs in an https sitemap).
Subpath sitemap (/catalog/sitemap.xml) may only list URLs under /catalog/.
List one canonical version of each URL (not both http and https).
No session IDs in URLs.
Include frameset and frame content URLs if site uses frames (legacy).

Next.js (App Router)

Use app/sitemap.ts exporting default async function returning MetadataRoute.Sitemap:

import type { MetadataRoute } from 'next'

const SITE = 'https://example.com'

export default async function sitemap(): Promise<MetadataRoute.Sitemap> {
  const staticPages: MetadataRoute.Sitemap = [
    { url: `${SITE}/`, lastModified: new Date(), priority: 1 },
    { url: `${SITE}/about`, lastModified: new Date(), priority: 0.8 },
  ]

  const posts = await fetchPostSlugs()
  const postPages = posts.map((slug) => ({
    url: `${SITE}/blog/${slug}`,
    lastModified: new Date(),
    priority: 0.7,
  }))

  return [...staticPages, ...postPages]
}

Next.js serves at /sitemap.xml automatically.
Use real lastModified from CMS when available.
For multiple sitemaps, use generateSitemaps (Next.js 13.3+) or route handlers.

robots.txt

User-agent: *
Allow: /

Sitemap: https://example.com/sitemap.xml

After publishing

Validate XML against sitemap.xsd.
Submit in Google Search Console / Bing Webmaster Tools.
Add Sitemap: line to robots.txt.
Monitor crawl stats; update lastmod when content actually changes.

vs JSON-LD and llms.txt

| Artifact | Role | |----------|------| | sitemap.xml | Enumerate indexable URLs for search crawlers | | JSON-LD (json-ld) | Structured semantics per page (rich results) | | llms.txt (llms-txt) | Curated map for AI agents (not exhaustive URL list) |

Use all three (sitemap, json-ld, llms-txt) for a complete discoverability stack.

Common mistakes

Wrong lastmod — sitemap generation date instead of page update date.
Relative URLs — must include https://.
Mixed hosts — subdomain URLs in www sitemap.
Session IDs in <loc>.
Duplicate http/https entries for same page.
Priority 1.0 everywhere — dilutes relative signaling.
Unescaped & in query strings — breaks XML.
Oversized file — split before hitting 50k / 50MB.

Related skills

json-ld — structured data on pages listed in the sitemap.
llms-txt — llms.txt for AI assistants (curated, not a sitemap replacement).

Additional resources

reference.md — Protocol, FAQ, schemas, Search Console, Next.js.
Protocol: https://www.sitemaps.org/protocol.html
FAQ: https://www.sitemaps.org/faq.php

Agent Skills: SEO — Sitemap

Install this agent skill to your local

Skill Files