Post

what happens when a notice is wrong? there's no clear path for challenging automated notifications with this after first read, just ignore?

also notices need to have verification that theyre coming from who they say they are

this part covers both doesnt it? Inter-service auth would be required: notices can not be submitted anonymously. Response error codes could indicate if the notice had been rejected.

challenging a result would be between you and your PDS operator (maybe also you?) - like if I operate email and one of my users gets blocked for spam reports, they appeal to me

i assume the notifications would be a "hey youve got ____ content" with the assumption that they might be labeled by the mod service (prolly manually configured) so you could use the standard appeals however that works

(will write up as a reply when I can) 1. 'comments' is fraught 2. type should leverage moderation infrastructure (require a label the service provides) 3. consider an authed request for updates/more info (maybe where comments go?) 4. consider allowing this pre-commit (github.com/bluesky-soci...)

1. always is 2. I feel like that conflates labeling and infra moderation, which we are trying to keep distinct. for labeling flexibility is a positive, I think here it is detrimental to interop 3. hrm that implies a more stateful interaction. this is designed to function more like an abuse@ email

i would love to also see this being used for things like relays sending spam notices to PDSs so relays can avoid permanently blocking entire PDSs just because of a singular spam account. would there be more specifics on the response to these notices that would allow things like that?

specifically responses that can more specifically indicate the action the notice receiving service has taken based on the notice