I can't actually seem to replicate the issue, sorry! I tried under Waterfox Classic 2021.02 as you mentioned, and it appears to work fine for me. Would you mind sending your settings (through the options->export button)? Contact links are available on the extension page (use whatever you're more comfortable with). Also, could you send an example link that doesn't work for you? For me, https://pbs.twimg.com/media/DlOC3mOWsAEi1ld.jpg redirects properly to https://pbs.twimg.com/media/DlOC3mOWsAEi1ld?format=jpg&name=orig .
It could be that there's a specific combination of settings that causes it to break, it wouldn't be the first time that has happened under Waterfox (due to its older JS engine).

Sorry, there's currently no way to add manual rules. This is something I'd really like to add in the future (see https://github.com/qsniyg/maxurl/issues/41), but currently there are a few technical problems with implementing this for the moment (see the linked issue for a bit more detail on why).

If you want to request a new site to be added to the script, feel free to use any of the contact links listed at the bottom of the script's description, whichever you're most comfortable with. I'm more than happy to add any new website to the script, and if I'm online, I usually add them within minutes of being notified :)

If you'd rather add it to the script yourself (note that this will require sideloading the addon), I've written a bit of a guide here: https://github.com/qsniyg/maxurl/blob/master/CONTRIBUTING.md#websiterule-contributions. In your case, a rule like this should work: if (domain === "subdomain.your-website.com") return src.replace(/(\/[0-9]+-)small(\.[^/.]+)(?:[?#].*)?$/, "$1original$2");

By the way, while the github issue hasn't seen much activity for a while, this does not mean it's been forgotten or ignored, the script is already getting restructured in many areas, which will facilitate the implementation of this. It's just that it is pretty a massive project (currently requires rewriting >70,000 lines of source code - much of which will be automatically handled, but even that requires a very significant amount of work).

Glad to know it's helpful!!
I should note however that this does not replace using a reverse image search engine, but rather complements using one. It'll find larger versions stored within the website itself (either through URL patterns or API calls), which are sometimes larger than what reverse image search engines can find (unless they crawl a page that contains the original image), but if the original image uploaded on the site is smaller than what's available on another site (e.g. an image reuploaded to a blog), it won't be able find that version.