Interesting approach. I’ve not thought about the implementation of actual logic yet.
Also, the difference between your bot project and my botnet is that your bots should act based on triggers (I guess) and my botnet is intended to be controlled by a human (of course, there is space for some custom triggers too).
The medium-term goal is to enable both responsive and proactive behaviors on the part of the bots. Figuring out the best implementation of that is the part we could vibe on I think 
Just got an email from Twitter:
Hello,
Your Twitter account (or accounts associated with you) currently has multiple registered applications, in potential violation of Twitter’s rules on the use of multiple API keys for a single use case. You can learn more about those rules here.
If these applications serve distinct use cases under our policies, please reply to this email with a specific use case for each registered application.
If these applications do not serve distinct use cases, please log into your existing account on https://apps.twitter.com/ and delete any applications which are in violation of our rules.
When you have completed these actions please reply to this email. Once we have confirmed that your existing access complies with our rules, we will continue our review of your application for Twitter Developer Account access.
Thank you for your interest in building on Twitter.
Regards,
I really wonder if they don’t have better things to do.
