Just to be clear… I have so far only tried fetching file:/// URLs from contentscripts running in “file: page content”. I have no reason to believe it works in “web page content”, but I haven’t tried it. I have also only tried fetching images already shown in the browser.
But yes. I will investigate a bit more and probably open a feature request.
PS. I have no current plans to release the MV3-version of xIFr any time soon, but if someone is curious about the “MV3 multi platform” support, I have checked it into a branch in my github(*). But be warned, it is not a little easy-read “demo extension” 
(though, if you focus only on how script in background is set up, that’s probably not too difficult to get).
(*) Sorry, but in the light of another current issue with spam-copies of my extension being posted on AMO, I don’t feel like linking directly to repository or the new branch right now.