TrustGrowth link audit ignores HTML <base> element and reports false broken links

The TrustGrowth audit for inetsoftware.de dated 2026-08-25 reports 58 broken internal links across 60 crawled pages. The reported targets return HTTP 404, but none of them are actually referenced by the current deployed HTML.

The affected pages contain:

 The crawler appears to resolve relative links against the page URL instead of honoring the document’s <base> element. For example, a valid link such as:

  href="en/i-net-cowork/download/"

  is incorrectly interpreted as:

  /en/i-net-cowork/en/i-net-cowork/download/

  Similar malformed targets include:

  - /en/i-net-cowork/en

  - /de/company/de

  - /en/en/i-net-helpdesk/download

  - /de/i-net-helpdesk/en/i-net-helpdesk/whitepaper

  Independent validation found:

  - 58/58 reported target URLs return HTTP 404.

  - 0/58 reported targets occur in the current deployed HTML.

  - All 58 source pages are reachable.

  - Current links resolve correctly when the <base> element is respected.

  Expected behavior:

  The crawler should resolve relative URLs according to the HTML standard, honoring <base href>. It should report only links that are actually present in the parsed document and resolve to an unreachable target.

  Impact:

  The audit produces a large number of false-positive critical issues and may lead users to make unnecessary Website changes.

  Suggested fix:

  Apply <base> resolution before normalizing and validating links. Add regression tests for documents containing a root-relative <base> and relative language/product links.