Add a page to machinebehavior.io

A new page passes the gate when it has a directory URL, a self canonical, a sitemap entry, an llms.txt line and root-absolute links. It joins the navigation with one entry in site/nav.yml. Docs pages get all of this from the generator; see Docs tree.

Steps

1. Create <name>/index.html. The URL is /<name>/.
2. In <head>, add <link rel="canonical" href="https://machinebehavior.io/<name>/"> and a meta description.
3. Link only with root-absolute paths: /man/, /style.css. Relative paths and links ending in .html fail the gate.
4. Add <url><loc>https://machinebehavior.io/<name>/</loc><lastmod>YYYY-MM-DD</lastmod></url> to sitemap.xml.
5. Add a line with the title, URL and a one-line summary to llms.txt.
6. Put at least one ISO date on the page; the freshness check reads the newest one.
7. Add the page to its section in site/nav.yml and run python3 scripts/site_chrome.py. It writes the top bar, the breadcrumbs, the sidebar and the footer, and reports the page as an orphan if the entry is missing.
8. Run the dry run (see Deploy pipeline), then pull, commit and push.

Links built in script

If a page builds links in JavaScript, set a.href in code. The gate reads href="..." in the page source, and a template string such as an interpolated href is read as a relative link. See Link check flags a template string.

Chrome and layout

Do not write a menu, breadcrumbs or a footer into the page; scripts/site_chrome.py writes them between <!-- mb:NAME --> markers. A research page links /style.css and puts its content in <div class="sheet"> (reading layout). An Inside page puts its content in <main> and keeps its own styles (app layout). A page that needs its breadcrumbs or footer somewhere else holds the empty markers there, for example <!-- mb:crumbs --><!-- /mb:crumbs -->. See Site navigation and chrome.