Methodology

How the claims on this site are produced, checked, and disclosed.

Content update cadence

Each section of the site declares how often a crawler should expect it to change, and the sitemap carries a last-modified date for every URL. Blog posts are declared daily, guidebooks and the homepage weekly, and the legal pages monthly.

Those dates are not decorative. A blog post's last-modified date is the post's own updated timestamp, and when a post has been edited more than a day after publication the change is also shown to you on the post itself, above the body. Guidebook chapters inherit their parent guidebook's updated timestamp, so editing a guidebook moves all of its chapters forward together.

Blog tag pages are deliberately kept out of both the sitemap and the search index: they repeat the content of the blog filtered by tag, and indexing them would put near-duplicate pages in front of you.

Affiliate disclosure

Every outbound company-logo link on the homepage is marked rel="external noopener noreferrer nofollow sponsored". The sponsored value is the strongest disclosure marker a link can carry, and it is applied to all of them without exception.

To be explicit about what that strip is and is not: it is framework attribution, not paid placement. The companies shown are public adopters of Ruby on Rails, Elixir Phoenix and SvelteKit — the frameworks these templates are built on. The heading above them says so directly: "Built on stacks that power these companies."

None of them is a customer of ThinkCodeShip, none has endorsed it, and no money has changed hands in either direction for their appearance. The links are marked sponsored because that is the conservative choice for a commercial page, not because a sponsorship exists.

AI policy

This site and the templates it sells are built with AI coding assistants. That is the product's subject, so pretending otherwise would be absurd. The interesting question is what stops an assistant from inventing things, and the answer here is machine enforcement rather than good intentions.

The repository carries rules that fail a test rather than a review. Testimonials, biographies and other content a human must supply are structurally unable to render until a human has actually written them — a half-written record fails its contract test instead of shipping. Social profile links are never inferred, so an unverified profile is absent rather than guessed. The product's structured data carries no star ratings or review counts, because no reviews have been collected and a self-awarded rating is a fabrication a search engine is right to penalise.

Where this page cannot yet state a procedure, it shows nothing instead of describing one that was never performed. Sections appear here as they become true.

Questions about any of this

Who maintains the site is on the about page. Contact details are in the footer of every page.