Set up antiraid

Antiraid handles waves of joins and raid-like activity from members. Antinuke handles destructive permission and moderation actions. Set them up separately; enabling one does not replace the other.

Before you start

You need the real Manage Server permission for antiraid configuration. Fake permissions do not unlock it. Flanker needs permissions for the response you select and access to the channels it protects. Keep its role above members it must moderate.

Choose a private alert channel staff can read. Captcha-based responses let suspicious members verify; kick and ban responses remove them and can also affect legitimate new members if configured too aggressively.

How a raid response works

Detect a pattern
Configured joins, chat activity or resource bursts cross their limits.
Contain the raid
The selected response can hold suspicious members behind verification and lock affected access.
Review and reopen
Staff review held members and end raid mode, restoring access through recovery.

1. Run guided setup

On the dashboard, select the server, open Antiraid and start setup. Beginner uses recommended limits. Advanced exposes individual detection rules and responses.

Or run:

/antiraid setup wizard
/antiraid log #antiraid-log
/antiraid config

The log command changes the destination. Replace the example channel with your staff channel.

2. Review joins and chat separately

Example join ruleExample values from the command reference. Adjust them to your normal join rate.

Join detection: On
Join limit: 8 joins in 10 seconds
Response: Captcha
LockdownThe --lock option controls lockdown for the join rule.: Review separately

/antiraid joins on --joins 8 --window 10s --do captcha

A busy event can generate legitimate bursts. Review normal join volume before tuning the rule. Avatar and account-age rules are separate checks; an account without a picture is not necessarily malicious.

Chat flood rules consider messages, participating accounts and repeated text. Media, repeated-link and Discord AutoMod correlation have their own limits. The AutoMod correlation rule needs an enabled Discord mention-spam rule; one blocked attempt alone does not qualify.

Raid mode only starts when two or more accounts flood together. A single account that floods is handled on its own: its messages are deleted and it gets the flood action (a captcha or jail). When the flood action is server lockdown, a single account is timed out for 30 minutes instead, and the server stays open.

Text chopped into short lines to slip past filters, like an invite split across lines, counts as risky. Three risky messages from one account within five seconds deletes them and restricts that account the same way. Chopped-up invites are also caught by the invite filter on their own.

See the full antiraid options for every rule, supported response and limit.

3. Protect tickets and temporary voice channels

Flanker’s ticket and Join to Create burst detection defaults to six distinct accounts in three seconds. Repeated attempts from a single account do not fill that distinct-account limit.

/antiraid tickets toggle on --accounts 6 --window 3s
/antiraid voice toggle on --accounts 6 --window 3s

During raid mode, Flanker stops new tickets and locks creator channels, voice joins and voice-channel chat. Existing rooms remain except recent rooms made by flagged raid accounts.

Another bot’s ticket panel

Use /antiraid tickets setup to provide its bot ID, panel message link and ticket category. Use a dedicated panel channel: raid mode hides that channel. Other apps’ button clicks are private; Flanker uses audit events and permission changes instead, and does not punish ambiguous openers.

Another bot’s voice creator

/antiraid voice add #join-to-create

Flanker’s own creators are included automatically. Add only external creators that need monitoring.

4. Review members and end raid mode

/antiraid unverified
/antiraid verify @member
/antiraid end

Review the held member before verifying them. End raid mode once staff have checked the situation. If a channel or role cannot be restored, read the recovery result and fix the specific permission or hierarchy problem instead of repeatedly toggling protection.

When raid mode ends, an incomplete channel restore shows Retry restoration instead of End raid. The remaining count refers to channel restoration, not an active raid. Fix Flanker’s channel access or Manage Channels permission, then retry. Once restoration completes, the alert has no recovery button.

Raid reports are retained for retry if delivery fails, independently of channel restoration. If the antiraid log is unavailable, restore its access or select a working log channel. Reports awaiting delivery survive a bot restart.

The manual /antiraid raid command is a moderation action: it acts on everyone who joined within the supplied duration and starts raid mode. Do not use it as a harmless test command.

Verify your setup

Confirm enabled rules and alert channel with /antiraid config.
Check Flanker can read, write and manage the relevant channels.
Explain verification and recovery to staff before an incident.
Confirm watched external panels and creators belong to the intended bots and categories.

Troubleshooting

No alerts: verify the log channel and channel overrides.

Legitimate joins are flagged: compare actual join volume with the rule, review account-age/avatar checks and adjust the relevant limit. Avoid broadly whitelisting people to hide an unexplained problem.

A member remains held: inspect /antiraid unverified, verify the intended member and review any permission-recovery errors.

Reset everything: /antiraid reset is owner-only. It unlocks channels and releases captcha members before resetting settings. Failed recovery records are kept. Team site admins have a separate dashboard reset exception.

Next: honeypot and welcome messages.