Element-web: Ability to preemptively ban users

[TrechNex]

Is your suggestion related to a problem? Please describe.
At the moment you can not ban a user until they have first joined a channel. This is problematic for two reasons:

a) If that user is harassing someone by DM for being a member of our project we can't do anything about it

b) If a project (or "community") has more than one channel then it becomes a game of "whack a mole"

Describe the solution you'd like
In the channel settings I would like to be able to search for a username and ban that user.

Failing that, at least some kind of "report" function for users who are unaware that they don't have to accept new chat requests from random strangers. This would not really solve the problem, but as a project admin I'll take what I can get! :D

Describe alternatives you've considered
Not using Matrix? I'm not sure what I need to fill in here.

Additional context
N/A - could link this issue to #5320 "ban users from groups", #6976 "shadow ban users" and #1866 "bulk kicks/bans"

[turt2live]

Use the /ban command:

[TrechNex]

@turt2live we'll give that a try when it next happens. I'd point out though that this solution was not obvious to us, and a GUI solution in Riot would be most welcome.

[e-lisa]

Hi, I think this is related to #4125

I also run a community. We have recently have had to shut down our Element Web service down until further notice due to the massive abuse.

Not being able to allow the chat admin staff to ban user's with out help of a sysadmin is extremely problematic, we have the exact same problems @TrechNex.

Also wanted to respond to @turt2live - That only works for a channel, an abusive user is still connected to the server and is able to do such activity such as:

• Harass other users by private messaging them
• Evade the ban by creating a new account with a disposable email
• Create their own rooms with vulgar/explicit name/descriptions

I just wanted to point this out that with out a feature similar to this, anti-abuse with Element is near impossible.

Training chat admin staff on how to call APIs by hand has a lot of pitfalls, in the heat of the moment the less steps required to ban a bad actor, the better.

[turt2live]

This issue only deals with rooms (or channels) - particular server behaviour and moderation is unfortunately a task for that server. We're well aware of why you would want to ban someone.

I suggest a read of https://matrix.org/docs/guides/moderation instead of posting similar complaints to multiple issues.

[TrechNex]

It would also be good if the ability to ban domains (i.e. "known bad" homeservers) was added in the same change, but I'm aware that might be better off as a different ticket.

(Currently you have to use Server ACLs, which basically means maintaining a JSON file that you periodically update and re-apply. That seems less than elegant)

[e-lisa]

@TrechNex If you want to see this implemented server-side I strongly suggest you up-vote and comment here: https://github.com/matrix-org/synapse/issues/121

@turt2live

I suggest a read of https://matrix.org/docs/guides/moderation instead of posting similar complaints to multiple issues.

The use-case discussed is not covered in the page you linked (and I have read it multiple times before coming here). These features are important and IMHO critical to the success of this platform as it grows.

There is an open ticket: https://github.com/matrix-org/synapse/issues/1216

If that feature request is merged, this ticket then becomes very do-able. It seems prudent to track them on both ends. Please let me know if this is the wrong thread, maybe the issues be merged/cross-referenced? Let me know.

[turt2live]

https://github.com/matrix-org/synapse/issues/1216 is completely unrelated to this issue and will not be merged. This issue is focusing on supplying a button somewhere in the UI to ban a known user before they are in the channel, as currently the only way to do so is through the /ban command.

IP bans and other forms of access control are out of scope for this issue.