While the placeholder tags have been removed, this project is still very much experimental - Dragon now works, but it doesn't always work well. Common sense while using is still required for use.
My revision history currently contains a lot of Dragon edits as well, with varying degrees of how extensive the edits are. All edits performed by Dragon in general are also labeled, meaning they can at the very least be searched using SEDE (with a delay of a week).
Dragon is a 2k tool largely meant to edit specific types of posts on Stack Overflow, because doing it manually is a ridiculous waste of time - by the time you edit one, there's two new ones to take its place.
I'm also being vague about "specific" because it doesn't automatically imply just one category. The current intended categories include posts containing thanks, as well as questions tagged [android-studio] if I can figure out a good metric for editing these.
For the near foreseeable future, editing is going to be supervised to verify that it does what it's meant to. All stages of the development process are going to be open-source as well - as a result, it's possible to use before it's properly done. Note that I advise caution if you decide to do so - this tool is going to be experimental for a while, meaning there could be bugs in how it operates.
This program is currently intended to be Linux-only (largely so I can use xdg-open for OAuth and just expect it to work without needing to go out of my way to test and verify) -- patches making it truly portable are welcome, however.
Issues, complaints, and ideas should be posted on GitHub, and not here, to better help me track what I've handled, and what I haven't
The code is written in Python, because there's no API libraries for C++ that don't use Boost and I can't be bothered to make one.
The code is available on GitHub, along with the specific requirements.
Dragon itself is now barely usable, though there's still substantial flaws in the various regex rules, and a ton of missing filters and/or edge-cases to existing filters that need to be accounted for to actually cover stuff.
Only manual edits are allowed at the moment, but it's possible to play around with. (But use common sense - it's also a good idea to double-check any edits). Automatic edits haven't been implemented because the regex hasn't been fully tested, and may cause damage at this point in time of run unsupervised. The goal is to get to a point where the regex is virtually bulletproof before edits are fully automated, but there's still a decent chunk of work required before I get there.