Link Signals sees only what a visitor or a search crawler can see. It does not log in, fill in forms or run your JavaScript. It does read the text of a page’s own inline scripts and decodes encoded strings in them, so links that those scripts write are found.
Links that are built only while a page runs, or that sit behind a login, are missed.
What is and is not seen
| Where the link is | Seen? | Why |
|---|---|---|
| In the page’s HTML | Yes | This is the main thing read. |
| Written by an inline script, in plain or encoded text | Yes | The script text is read and decoded, not run. See script-written links. |
| Written by a script loaded from another file | No | Only inline script text is read. |
| Added by a script after the page loads | No | Nothing is executed, so nothing is assembled. |
| Behind a login, such as a members area or a staging copy | No | Link Signals never signs in. |
| In a comment held for moderation | No | Only what is on the page when it is read. |
| Shown only to search crawlers | Partly | A sample of pages is also requested as a crawler would see them. See below. |
| Hidden by a stylesheet class | Not as hidden | Only inline styles are recognised as hiding a link. |
Does it pretend to be Google?
For one comparison only. On the homepage and about 10% of pages, Link Signals also requests the page with a Googlebot-style user agent and as a visitor arriving from a Google search, and flags links that only those views show. The requests come from its own servers, not from Google’s network, so a site that verifies a crawler by its address can tell the difference and may show nothing unusual. Cloaked links explains what this finds and what it cannot.
Why a link may not show in your browser
A cloaked link is not served to you on purpose. If a report lists a link you cannot see, open the page named under Found on and look at the file, theme or plugin that adds it, rather than the page as you view it.
What the report shows from the page
The report keeps the first 120 characters of the text inside the link tag, with the spacing tidied. A link built from an image shows the image’s alt text. To judge whether a link belongs, Link Signals uses the title, first heading and meta description of the page it is on, with the anchor text and the destination, not the whole page.
Where to read more
- Accuracy and limits: every blind spot, in one table.
- Hidden links: what is detected and how it is shown.
- After a WordPress hack: what to do alongside a scan.
- Google on JavaScript and search: how a search engine handles script-built content.
Questions
Can I scan a staging site that needs a password?
No. Because the pages need a login, they cannot be read. Scan the public address instead, and remember each address has its own history.
Will a cookie banner or consent wall stop it reading a page?
Not usually. When a link destination bounces a first visit through a consent address, redirects are followed and cookies the site sets are kept. For your own pages, the content that is in the HTML is read whether or not a banner is shown.
Could a link I added with a widget be missed?
If the widget builds the link in the browser from a separate script file, yes. If it prints the link into the page’s HTML, or in an inline script, it is found. Viewing the page source and searching for the destination shows which.
Run a scan on a public sitemap and compare the report with your own page source.
