Rubric: link integrity

Written by a person. Last read by a person on 2026-09-04, 4 days ago. Its facts were checked by the eval suite on 2026-09-07.

Enforced by check_links.

The claim

Every internal link resolves, including the prerequisites and next steps that carry a reader through the documentation.

Observable evidence

• Each url in a page's prerequisites and next_steps frontmatter.
• Each root-relative markdown link in the body.
• The set of pages the content tree actually produces.

The pass bar

Every internal target resolves to a page that exists. A fragment is checked as far as its page.

Prerequisites and next steps are held to the same bar as body links rather than a lower one, because they are the navigation a lost reader falls back on, and a broken link there fails precisely the reader who most needed it to work.

Ambiguous cases

External links. Not checked. Checking them makes the suite depend on the reachability of sites this project does not control, so a green build would mean the internet was up. External link rot is real and is a gap this rubric does not close.

Fragments. The page is verified; the anchor within it is not. A link to a section that has been renamed will pass. Closing this needs the built HTML rather than the source, and is worth doing when deep-link orientation lands in change 0004.

Links to pages a later change will add. Fail, deliberately. A link to a page that does not exist yet is broken now, whatever the plan is, and a reader following it today gets nothing.