It is working for me, I tried to find ways to break it and I haven't been able to find anything other than it doesn't try and download tiddlers that use _canonical_uri fields (you can just download them directly from their location, I may update it to download them in the future).
It shouldn't be limited to working with Bob or node, the behaviour is completely in the browser.
There may be some browser setting that prevents it, but I am not aware of anything that would do that.
The types it downloads are limited to 'image/gif', 'image/x-icon', 'image/jpeg', 'image/jpeg', 'image/png', 'image/svg+xml', 'application/pdf', 'application/zip', 'application/font-woff', 'application/x-font-ttf', 'audio/ogg', 'video/mp4', 'audio/mp3', 'audio/mp4' because those are the ones I thought of at the time. It won't work if the tiddler doesn't have one of those types. Eventually I will probably add a setting that lets you change the list of mime types it will download.
It also will not try and download images that are included using _canonical_uri, the plugin will take tiddlers that are base64 encoded and try to download the non-base64 encoded version.