There is nothing inherent to Beta (yet) that makes it work any different from Release when it comes to interacting with the client. Both versions use the same exact underlying code, so there was nothing added or changed lately that would affect that.
Beta #863 contained mostly API additions to things that we were manually updating each time, and it was slowing us down too much. All of the new Dat stuff is the same exact things we were using before, it's just we dumped the files from disk rather than memory, and then copy/pasted stuff into online links for users to reference. The new setup allows users to obtain that info from the API directly, which is much better. If anything broke as a result of these changes, which none of the new code except for one aspect of it is actively used in the current Beta, it'd be pretty noticeable, as the bot would not be able to take waypoints or area transitions at all. I did checks though all 3 difficulties and acts, and did not notice anything strange, so I think that's highly unlikely.
The only other significant change in #863 was a bug fix to LokiPoe.GuiControlSize. If the bug was not fixed, it would be possible for the bot to try and interact with the 'Resurrect' button at the wrong location, and never be clicked. I don't think this ever was an issue, because if it were, users would report the bot fails to res, which hasn't happened yet. Nothing else uses that functionality in the API in the current version you guys are using.
Other than that, there was nothing else that changed that would affect stability or whether or not you get flagged, so I think the flag was just coincidental. It's happened before though, people have upgraded to the next version, and at a close time after, they get flagged or banned, and think it was that version, but in reality, nothing changed to affect it. It's one of the nuisance of server sided detection, you can't really know what triggered it. However, I will say, I do make new accounts on different IPs/MACs and regularly test Beta to make sure there isn't anything changed that seems to cause a "fast flag". So far, I've not had any of the accounts I've test flagged or banned, so I've not seen anything yet that would cause me to think otherwise. That is not to say some of the bot logic could be improved some more, but that's why the plugins like StuckDetection, Scheduler, etc.. were added as some base things people can use or modify if they are botting for longer periods of time.
As for the issue as a whole, bans/flags will happen. GGG's server sided detection is always evolving to detect new patterns of botting, and as a result, bot logic has to also evolve to keep up. The new Beta is a good example of why things are being rewritten, so people have more customization and the ability to do what they need to try and avoid some issues that come along with "public bots". I do expect they are gearing up with new stuff for the big 1.2 update, as it's getting a lot of attention. As usual, anything we can do to help users avoid flags and bans, we'll take seriously and change stuff around if needed.
Beta is about to get another very large update in less than a few weeks (separate from the 1.2 update stuff) that does change how key parts of the API work with the client. The idea is to hopefully avoid some of the long time input issues and other various things that still seem to happen. This change should also lower maintenance time version to version, so that's why there has not been much new stuff the past few days. It's very important this stuff gets done as soon as possible.
As a result, pretty much all of the current Beta stuff is being rewritten again for this new API compatibility. There's many reasons for this additional rewrite, including potentially less downtime for 1.2, but hopefully it does help address some of the things the bot does that users worry about potentially causing flags (which I can't say they actually do, but user confidence in the bot is very important).