Add a more concrete way to determine image on page

Sometimes a link to an image might seem correct. But the application flags of the content of the link as none image. We might have to look for a better way than just reading the content-type from the headers returned

No files, tests, or entry points are named. Start by locating the Python logic that determines whether a page link contains an image and currently relies on response headers; reproduce a case where the content type is reported as non-image, then define completion as reliably identifying valid image links without that false negative.

Written by the indexing model from the issue text.