Skills· Official
SEO
The foundations that get a site crawled, indexed, and shown well — per-page metadata, one canonical URL, a sitemap, honest robots rules, and structured data.
When to use it. When launching a site, or when pages aren't getting indexed or are showing the wrong title in search results.
Search engines reward sites that are easy to crawl and clearly labelled. Get the foundations right before chasing anything clever.
One page, one address, one title
- Every page has its own
<title>and meta description — written for a person skimming results, specific, not boilerplate. Keep titles under ~70 characters and descriptions ~70–160. - Each page declares one canonical URL. If the same content answers on two hosts, both name the same canonical. Duplicate URLs split your ranking.
Tell crawlers the shape of the site
- A sitemap.xml listing every real page, kept current.
- A robots.txt that allows what should be indexed and disallows only what shouldn't (settings, APIs, sign-in plumbing). Don't accidentally block the whole site.
Help the result look right
- Structured data (JSON-LD) for the things search engines render specially: organisation, FAQ, product/offer, breadcrumbs. Keep it truthful and in sync with the visible page.
- A share image and Open Graph/Twitter tags so a shared link looks intentional.
Make it fast and stable
- Serve pages quickly (cache static content at the edge; avoid per-request work on public pages). Speed and stable layout are ranking factors and a better experience.
Verify, don't assume
Fetch the live pages and check the actual titles, canonicals, sitemap entries, and robots rules — build-time intent and what's served can differ. Then submit the sitemap in Search Console and watch what gets indexed.
Use this skill
Skills work with the coding agent in your Gaitro project — add one and the agent uses it when the work calls for it.
More skills
This skill as markdown, for your agent: /skills/seo.md.