Conversation

Notices

1. Embed this notice
   BeAware (beaware@social.beaware.live)'s status on Wednesday, 09-Oct-2024 04:36:33 JST BeAware

   It sure would be nice to have some automated reporting by keywords right about now....you know, the thing I was explicitly told was NOT gonna happen...🤬

   #Spam #Nonsense #Moderation #Mastodon #Fedi #Fediverse #FediAdmin #MastoAdmin

   • Embed this notice
     abff08f4813c (abff08f4813c@j4vcdedmiokf56h3ho4t62mlku.srv.us)'s status on Wednesday, 09-Oct-2024 07:48:56 JST abff08f4813c

     in reply to

     Thinking it over, it would be very easy to do this actually.

     Not saying it's a good idea, just that it's technically possible - and made easier by the technical details of federation.

     A modified, rogue instance could scan each post/comment/toot as it federates in against a list of keywords, and use a bot account to report the ones that fail the scan back to the original instance.

     Worse, this rogue instance might not mark the account as a bot.

     BeAware repeated this.
   • Embed this notice
     abff08f4813c (abff08f4813c@j4vcdedmiokf56h3ho4t62mlku.srv.us)'s status on Wednesday, 09-Oct-2024 10:20:21 JST abff08f4813c

     Ah, sounds like you basically want the opposite of https://github.com/db0/threativore/tree/main/threativore - this bot finds things based on regex matching, and then "reports" it - i.e. report as spam to the moderators/admins. You instead would want it to take a different action - I guess retooting it to you (or I suppose it could also email you a link to the toot).

     I think, and keep in mind I haven't ested this, but for a dev, it seems like this is still pretty easy to do by modifying threativore - just change this line here, https://github.com/db0/threativore/blob/main/threativore/threativore.py#L339 from the moderation-reporting call to the "notify me about interesting posts" call.