What is actually inspected
The checker requests the public source page through a guarded server-side fetch, parses anchor elements in the returned HTML, and resolves relative href values against the source URL. It then compares those resolved URLs with the exact target after removing only the fragment and normalising default ports.
The result records matching anchor text and rel values such as nofollow, sponsored, or ugc. Those are observations from one response, not a historical record and not a statement about how every crawler interprets the page.