2 suggestions regarding removing bots from servers

[Andre601]

This discussion will have 2 ideas which both are related to removing a bot from a Server.

1) Display Remove Bot in context menu instead of Kick <user> and Ban <user>

When removing a bot is it currently through either a kick or ban.
Especially the ban seems a bit pointless given that a new invitation of the bot through its OAuth2 link nullifies this ban, so a kick should be the go-to option here.

This suggestion is more about the Client-App and its appearance so it would be nice if that could be forwarded to the responsible team.

The idea is to remove the Kick <user> and Ban <user> context actions when right-clicking a bot and instead display a simple Remove Bot option, which in itself would just be a slightly changed version of the kick behaviour.
It would essentially open the same popup where you can set a reason for removing the bot, which would init a kick event in the process.

tl;dr Redesign the kick option for bots and hide the ban option as it's not any more effective than kick.

Here's a quick Mockup I made (Ignore the red line. It's from an embed):

2) Add a type and optional reason field to the GUILD_DELETE event

Right now do devs have no real idea why their bot has been removed from a server. It could've been a kick or the Server could've been deleted, you don't really know.

This is why I want to suggest to add a type field, which tells you what kind of delete event this was (KICK, BAN, DELETE, ...) and a optional reason field which, when available, would give the reason provided by the user. The latter field would therefore be used on a kick or ban of the bot and contain the kick/ban-reason when one was set.

Here are some quick example JSONs I made:

• Server has been deleted

  {
    "id": "41771983423143937",
    "type": "DELETE"
  }

• Bot has been kicked (No reason)

  {
    "id": "41771983423143937",
    "type": "KICK"
  }

• Bot has been kicked (With reason)

  {
    "id": "41771983423143937",
    "type": "KICK",
    "reason": "Bot stopped responding to any commands"
  }

The info could help developers discover bugs or server/shards where the bot is no longer responding properly, or to simply improve features.

[LagClear]

Great ideas hoping this gets added!

[MoonlightCapital]

I second this idea completely. I'd love hearing passively if my bot is malfunctioning or users have anything to share to me.

[Mehgugs]

I think the first idea is a good change and definitely makes more sense for integrations without a bot user attached (read slash command only bots).

Regarding the second change, I believe knowing your bot was kicked/knowing why you're no longer in the server could be a slight issue and could enable some undesirable behaviour. I think if you want to be able to know why your bot was removed for feedback purposes, the bot would need to support a command that accepts a reason and then leaves the guild.

I appreciate that you probably don't agree knowing kick reasons is bad but just consider the api sometimes has to ackowledge the lowest common denominator of developer exists, even if most people would not use that information for purposes that could be considered abusive.

[Andre601]

I dislike this "always go out of the worst dev existing" mentality.
Sure, bad actors exist, but always planning the API to not give them advantages for abuse is a really bad aproach since it also causes us good devs to constantly live with those disadvantages.

I see no real harm here in providing a optional kick reason. Many people probably won't give a reason anyway.

[Mehgugs]

As I said you can provide a /remove command to let people provide an optional reason should they wish to inform the bot and its developer. I am admittedly somewhat cynical, but I can easily imagine people making bots dm you a bunch of feedback related stuff unsolicited -- should they share a mutual server of course -- if they could see you kicked the bot. The reason itself is irrelevant mostly, it's the fact you can see you were kicked specifically.

[Andre601]

As I said you can provide a /remove command to let people provide an optional reason should they wish to inform the bot and its developer.

Why would they run a command if they can just right-click name -> kick <user>? Barely anyone would ever use this since the kick and ban feature have been become a major part in "removing" a bot from a server, so unless you would directly tell them to use that command instead to give feedback would it not have any real impact.

I am admittedly somewhat cynical, but I can easily imagine people making bots dm you a bunch of feedback related stuff unsolicited -- should they share a mutual server of course -- if they could see you kicked the bot.

And I again say that having features limited or reduced to a bare-minimum setup only to prevent abuse would not help anyone.
Discord already offers a lot of things that have the potential of abuse, such as sending messages itself.
They did make changes in the past to make abuse harder such as increasing the rate limit to a higher delay for renaming channels because it was abused by bots showing stats on VC names, yet here we are still having those bots doing it.
Or bots that update their status more frequently than what Discord normally tolerates.

Pretty much everything on Discord has the potential of abuse because it's there, it's free to use and has barely any real safety-guards outside of rate limits and authentication.

And while bad actors exist, are they not the majority here. They are only the more prominent ones to notice because bad news shares better than good ones.

[ChrisTalman]

I can easily imagine people making bots dm you a bunch of feedback related stuff unsolicited -- should they share a mutual server of course -- if they could see you kicked the bot.

Bots can already do this, though.

They can easily keep track of the servers they're in, including receiving an event when they have left a server, which is either the result of a kick, deletion, or their own choice to leave.

They can't see who kicked them, but they don't need to, if they want to be annoying. They'll know who has interacted with the bot, and could DM them if they have the opportunity. The person who added the bot will likely have interacted with the bot, so the bot could have their user ID.

In addition, I'd note that including the guild ID in an event is not integral to this idea. Although I don't think there is sufficient risk of abuse to do this, the type and reason data could instead be made available at a new endpoint, which bots could poll, and the contents of which could be delayed (not available as soon as the bot is kicked) to make it difficult to match the data up with specific servers. The data would essentially be anonymised.

[erkinalp]

Add a type and optional reason field to the GUILD_DELETE event

The human user client would benefit from that too, allowing to show users a message like "You have been banned from {guild.name} for: \n {reason}:".