cjslepHi, I would be curious if others had thought about whether it was possible via ActivityPub inbox forwarding abuse to cause victim servers to spam/DDOS servers (background: https://github.com/w3c/activitypub/issues/295)
cjslepReason being: while there is recursive depth protection for looking for linked objects owned by the server, there is no mention in the spec about protecting from too-wide to/cc/audience fields
cjslepAnd what also makes me wonder if it had been thought through is that Example 16 is actually one criterion away from a small scale version of this attack vector
saranixcjslep: short answer, no. Your server implementation should not do forwards that do not have permission. The discrete permission for this inhubzilla for example is "can forward to all my channel contacts"
saranixso in your example of popularperson1, popularperson2, etc., only if all of those people gave you permission to post to their channel (like a "forum") will you be able to
cjslepOK, thanks saranix. That gives me some ideas of how I want to proceed with my own lib impl. I'm not familiar with the inner workings of mastodon so I can't elaborate there.