Feature/#95: Prevent channels being misused for bot testing

[louis-young]

PR for #95

This pull request helps to prevent channels from being misused for bot testing.

Changes

• Add a list of channels in which the bot is disabled.
• Send a direct message to the user prompting them to test bot commands in direct messages.
  • If the user has direct messages from other users disabled on the server, notify them via the channel they triggered the bot command in.

Notes

• This is my first time ever looking at a Discord bot so please review with extra caution.
• I had to cast the channel object to a TextChannel to access the name property. I don't like this, but this appears to be standard from what I could find on some Discord.js issues.

[nickserv]

I understand the issue and I appreciate the effort, but I think that entirely disabling all bot commands in some channels restricts the usefulness of the bot too much. For example the !gender command is meant to be used as a response to a user misgendering, which can happen in any channel, including random. If you'd like to try, it would be nice to also have a list of commands that specifically get restricted.

[louis-young]

I understand the issue and I appreciate the effort, but I think that entirely disabling all bot commands in some channels restricts the usefulness of the bot too much. For example the !gender command is meant to be used as a response to a user misgendering, which can happen in any channel, including random. If you'd like to try, it would be nice to also have a list of commands that specifically get restricted.

That should be pretty trivial to do, I’ll take a stab in the morning 🙂

[awareness481]

Is accepting the communication commands enough?

Something like this should work imo:

const whitelistedCommands = ["!gender, !xy"];

export const isWhitelistedCommand = (commands: string[]) => {
  return (
    commands.filter(command => whitelistedCommands.includes(command)).length > 0
  );
};
//...
const hasPreventedBotResponse = isBotDisabledInChannel(channelName);
const isNotWhitelisted = !isWhitelistedCommand(command.words)

if (hasPreventedBotResponse && isNotWhitelisted) {
  notifyUserBotResponseWasPrevented(msg, channelName);

  return;
}

[vcarl]

Maybe a different approach would be to change how the cooldown works? Right now it's hard coded with a specific duration, we could make it an exponential backoff situation so folks get blocked for longer times if they keep spamming them, with a DM'd message like "it definitely works, you can DM here with no cooldown if you want"

[vcarl]

I'm going to close this out, since it's pretty stale (sorry!) and I think a different approach would be more successful. Let me know if you want to talk it through more, I like the spirit of this but don't think this is mergable as-is