Given the existing infrastructure behind the (1) first-time chat callout experiment, (2) non-mod chat delay, and (3) automod caught message queue:

There should be a toggleable setting where the initial message of any newcomer to chat must be approved by a moderator for that and any subsequent messages to appear to non-mods. Thus, this would limit the ability of newcomers to send messages without prior approval.

Not only would this reduce how many eyes see hateful messages from human accounts, it would neutralize spambots from having their messages show up to non-mods -- a useful step in responding to the current wave of "hate raids," especially when these messages can include streamer dox info.

Technical note: such a feature must work whether the incoming chat originated from a raw websocket or the first-party site. That is, this should not depend on a gql token negotiation system as is implemented for the new_chatter ritual.

Additional nice-to-haves:

• Bulk ban/allow buttons for the queue this would create: Bulk ban would be useful upon a spambot attack while bulk allow would be useful upon a legitimate raid

• Official pubsub topic for listening to these caught messages. Official helix endpoint for approving them. This would allow third-party chat clients (e.g., Chatterino, Chatty) to hook into this system for power mods. Also, this would allow specialized bots to automate actions (e.g., ban all first-time chat messages that match a certain regex). Note: this does not need to be implemented for the initial release of this feature request - it's more important to have an agile response (and some third-party devs can live with unofficial pubsub to get the job done)