As for spam prevention...we're still discussing. We obviously don't have the ability to study message contents. At the very least, we will need to have easy/clear blocking features
This immediately jumped out at me as a potential problem with the system. I think you had better get jumping on it quickly. Preventing abusive messages is much easier if mitigation is built in from the start.
What prevents hundreds of spam messages hitting a new user as soon as they sign up? An automated system could create a spam message for every github user account (just as a way to get likely signup identities) and keep a client running and they would be delivered to new users as soon as someone verifies themselves. A group of people could target one individual so that they are flooded with unwanted messages upon joining.
This immediately jumped out at me as a potential problem with the system. I think you had better get jumping on it quickly. Preventing abusive messages is much easier if mitigation is built in from the start.
What prevents hundreds of spam messages hitting a new user as soon as they sign up? An automated system could create a spam message for every github user account (just as a way to get likely signup identities) and keep a client running and they would be delivered to new users as soon as someone verifies themselves. A group of people could target one individual so that they are flooded with unwanted messages upon joining.