the-draupnir-project / Draupnir

A Matrix moderation bot
https://the-draupnir-project.github.io/draupnir-documentation/
Academic Free License v3.0
68 stars 13 forks source link

Replace `FirstMessageIsImage` with something more lenient. #467

Open Gnuxie opened 1 week ago

Gnuxie commented 1 week ago

Banning the user has always been too extreme and the protection is easily worked around. Needs replacing with configurable consequence providers for the action against the user and something that can take into account metrics about the user https://github.com/the-draupnir-project/Draupnir/issues/466.

The question then is whether we are just designing a protection that provides a permission layer ontop of Matrix that provides them in a configurable way based on some simple metrics. (Currently thinking how long they've been in the community, how much they've interacted (which i guess also depends on when those interactions took place and how many, and all of these probably need to be combined into one score, who knows).

It probably also makes sense for this protection to include the permission to send a link.

Mods could also possibly set the threshold for different consequences based on the metrics that we're thinking about with https://github.com/the-draupnir-project/Draupnir/issues/466. So if a link is sent on anything below 3, we'd redact but maybe above we'd alert the mods. That kind of thing.