Server Moderation (AutoModerator)¶
GiselleBot offers an auto moderation feature to be used alongside normal, manual moderation. The current auto moderator currently supports 8 triggers (messages or actions performed by users) and 5 actions (actions performed on the offending user and/or message). Each trigger can be configured with an extra whitelist, as described below.
This guide assumes you are familiar with the manual moderation module of GiselleBot. If you have any doubt about one or more of the actions or parameters that are used within the AutoModerator module, try checking Server Moderation (Manual) first.
By default, administrators and moderators (refer to !modrole) are immune to auto moderation triggers.
Server Invites: recognizes Discord server invites in user messages; this trigger supports shortened URLs (e.g. Discord invites hidden behind a bit.ly/gisl.eu shortening service), and ignores invites pointing to the current server.
Mass Mention: counts the number of mentions (roles, users or everyone/here) sent by a user, in a channel, within a certain span of time and triggers if the number of mentions is over a threshold. The default (allowed) threshold pair is 5 mentions in 30 seconds, but can be configured in each server.
Banned Words: checks the message against a list of words, configured by the user, and triggers if one or more words are found within the message. Punctuation and letter case are ignored. The parser can be configured to trigger on an “exact match” (e.g. banned word:
test, matching word:
test), if the banned word is found at the “beginning of a word” within the message (e.g. banned word:
test, matching word:
testing), or “anywhere in word” (e.g. banned word:
test, matching word:
Anti-Spam: counts the number of messages with the same content sent by a user, in a channel, within a certain span of time and triggers if the number of identical message is over a threshold. The default (allowed) threshold pair is 3 messages in 10 seconds, but can be configured in each server.
Anti-Attachment Spam: counts the number of attachments sent by a user, in a channel, within a certain span of time and triggers if the number of attachments is over a threshold. The default (allowed) threshold pair is 5 attachments in 10 seconds, but can be configured in each server.
NSFW Images: recognizes possible NSFW images sent by posting URLs in a message, or using message attachments, and triggers if at least one of the images posted is over the NSFW threshold for the server. Refer to NSFW Images Detection Tools for a deeper explanation of this detection system, and to !nsfwthreshold to configure the server threshold.
QR Codes: recognizes QR codes contained in images sent by posting URLs in a message, or using message attachments, and triggers if at least one of the images posted is confirmed to contain a QR code. This check works on both screenshots and pictures (taken by a camera/webcam), even though the recognition rate of pictures may be affected by the quality of the image.
Anti-Raid: counts the number of users (either new, or existing users leaving and re-joining) joining your server within a certain span of time and triggers if the number server joins is over a threshold. The default (allowed) threshold pair is 5 users in 15 seconds, but can be configured in each server.
Discord lag or connection problems can cause Anti-Spam and Anti-Attachment Spam false positives.
The Server Invites, Banned Words, NSFW Images and QR Codes triggers also support message editing: messages will be re-checked upon being edited.
The NSFW Images trigger, by no means, is supposed to reliably recognize all NSFW images. Use it at your own risk, and only as an additional tool to support humans in better moderating the server.
When more then one trigger is active and a message is potentially breaking more than one of the active triggers, this priority order will apply:
Server Invites > Mass Mentions > Banned Words > Anti-Spam > Anti-Attachment Spam > QR Codes > NSFW Images.
Automatic deletion of the offending message.
Auto-warn: Automatically apply a generic warning on the offending user. Specify a rule with the dedicated option.
Auto-mute: Automatically mute the offending user. The applied mute is a temporary, 2 hours long, mute.
Auto-kick: Automatically kick the offending user.
Auto-ban: Automatically ban the offending user.
When more then one action is configured on a trigger, this priority order will apply:
Ban > Mute > Kick > Warn.
Users: Ignore messages/actions performed by specific users in a server.
Roles: Ignore messages/actions performed by specific roles in a server.
Channels: Ignore messages sent in specific channels in a server.
Servers (Server Invites only): Ignore Discord server invites pointing to a specific server. You need to use the server ID to add this kind of whitelisting option. Applying this whitelist rule enables instant, temporary or permanent invites (including vanity URLs) for one or more server(s).
When adding or removing roles and channels through !automodsetup, other than toggling individual roles and channels, you can quickly add or remove all roles and/or channels from the list by using the following special tags: ALL_ROLES, ALL_CHANNELS, NO_ROLES, NO_CHANNELS.
Moderators alerting: Each auto moderator action will be logged into the Moderation logger (refer to Server Activity Logging). If this option is enabled, each log entry will also include a mention to the current moderator role(s).
Moderation rule: If a moderation action is taken against the offending user, this option will let you select one rule to use for that action.
Configuration of the auto moderation feature is achieved by using the following command. It will open an interactive menu within the current channel, using which you’ll be able to setup the module.
You must save the changes you applied (option 1 of the menu) in order for them to be applied.
The AutoModerator will also be configurable through the online dashboard, as soon as it’s available for public use.
Opens the auto moderation interactive setup menu. Use the menu items to configure the above settings.
Not all of the settings will have a meaning in all of the triggers. Read the above descriptions to understand what each option means within the specific trigger.