actually, this is a bug on our part where we're not parsing the correct parts of the document for image links. i'm pushing out a patch release now to fix this!