taut

🚧 This project is maintenance mode! 🚧

I will be fixing and responding to pull requests and issues, but it is not in active development.

What is this?

Taut is a self-bot for Slack. That means it can act like you and in many cases can't be distinguished from a human. This specific bot is primarily designed to act on triggers and impersonate you, but it also has various other features outlined below.

Have any questions? Feel free to create an issue on GitHub.

Setup

Currently, you'll need Node.js and Git installed to run Taut. If anyone else wants to use it I'll consider making it easier to run.

$ git clone https://github.com/kognise/taut.git
$ cd taut/
$ yarn

Next create the file config.js. This is the main configuration file. Unless you're very technically I recommend pasting the contents of default-config.js in for some default values.

Go to Slack's legacy tokens page and generate tokens for all the workspaces you want to use Taut in. Then paste the tokens one by one into the tokens array in the config, like so:

module.exports = {
  tokens: [
    'xoxp-etcetera'
  ],

  // ...
}

Now run yarn start to start the bot! If all went well everything should be working.

Commands

Right now Taut only supports 5 commands, contributions are welcome. Arguments in square brackets are optional and arguments in angle brackets are required.

Command	Description	Aliases
stats	Get info about the bot and current workspace	N/A
dox [user]	Get the profile of either the mentioned user or yourself	uinfo, info, user
pardon	Clears the suspicion score of a user, if you're the bot owner	N/A
coronavirus	Get statistics on the new Coronavirus (2019-nCoV)	corona, coronastats, coronastatus

There used to be a command to get your suspicion score, that's been removed due to excessive spam.

Spam Prevention

This bot is primarily designed to pretend to be real person which means it needs strong spam prevention to stay realistic and not be a nuisance.

Taut will analyze every message and use an algorithm to generate a score for each user on how likely they are to be trying to spam with the bot. By default Taut will automatically "shadowban" (silently stop responding to) users that exceed a certain score threshold.

You can update the threshold with the configuration value shadowbanThreshold, or set it to -1 to disable shadowbanning. Scores will be reset generally around 2 minutes which is often enough for people to cool down.

Taut can also ratelimit itself, the minimum amount of time between bot messages can be adjusted with timeout in the config.

Automation

As I've mentioned before, Taut's primary purpose is automation. It therefore has a hugely powerful system for pretending to you which might take some getting used to. Lo and behold...

Writing Automation Rules

Automation rules go under automations in the config file. It's an array of JavaScript objects. Here's a simple rule that replies with 👀 anytime someone sends exactly 👀:

module.exports = {
  // ...

  automations: [
    {
      match: /^:eyes:$/,
      response: ':eyes:'
    }
  ]
}

The match field takes a regex, or regular expression. If you aren't familiar, I recommend learning regexes and testing out your own regexes. A tip that might come in handy - if you want a regex to not be case-sensitive just put i after the last slash, like /regex/i.

Vary Your Responses

The automation I've shown so far has a response field, which is fine for most automations! But if you want to be more realistic, you might want Taut to pick a random response from a list, this is where responses comes in.

Below is a rule that'll reply to greetings (hello, hi, etc.) with a random response from a list of responses. It's fine if you don't understand the regex!

module.exports = {
  // ...

  automations: [
    {
      match: /^(he[lw]{2}o|hey+|hi+|yo|oy|sup)$/i,
      responses: [ 'heyo!', 'hey!', 'hiii', 'oy', ':wave:' ]
    }
  ]
}

For each automation you can also specify a probability field between 0 and 1 that will decide how often it shows up. This is helpful if a command is being too spammy.

Icky Channels

Some of your automations might not be appropriate in channels, or you might want to ban some channels altogether! This is where two new configuration options come in.

Per-rule, you can set rule.blacklistedChannels to a list of blacklisted channel ids.

For channels you want to block for all automations, you can add them to blacklistedAutomationChannels in the root of the config. To get a channel id, right click on a channel and choose Copy Link, then remove everything before the last segment.

More Configuration

Here's a quick run-down of other config options I might've missed:

• name sets the name of your bot
• prefix sets the prefix for commands, I recommend setting it to something different from other bots
• color sets the left border color for embeds
• statsFile sets the file to save message and user stats, don't bother changing this

Credits

Right now I'm the only contributor, although as I've mentioned a million times contributions are welcome. Shoutout to Max for the original inspiration with his selfbot Slacky.