Discord Anti-Nuke Bot: Stop Mass Bans and Channel Deletion

The attack it is built for

Spam filters watch ordinary members. A nuke comes from someone who already has permissions: a compromised moderator, a stolen admin account, or a bot you invited and should not have. In under a minute they can delete every channel and ban half your members. Discord does not stop this on its own, because every single action is allowed for that role.

Anti-nuke works on the pattern instead. One person doing many destructive things quickly is not normal staff behavior.

What it watches

Noxyr reads the audit log to find out who did each action, then counts per person:

• Channel deletions
• Role deletions
• Member bans
• Member kicks

When one person passes the limit inside the window, the bot punishes that person and posts to the log channel.

Settings and defaults

Set these with /security antinuke or on the Security page of the dashboard, and run /security status to see what is active. Starting with the action set to none is a safe way to watch what it would catch before it punishes anyone.

Trust list

Some people really do delete lots of things, such as an admin doing a clean-up. Use /security trust with a role or user to exempt them, and /security untrust to remove them. The server owner is always exempt and the bot ignores actions it performed itself.

Keep the trust list short: your bulk-cleanup tools and your most trusted admin. The trust list is never copied by Copy Settings, so a second server never inherits it by accident.

Where it can fall short

This is the honest part of any anti-nuke setup.

• The bot needs View Audit Log. Without it, it cannot tell who acted.
• To time out, kick or ban, it needs that permission and a role above the person it punishes. It cannot act on someone whose role is above its own.
• Discord's audit log can lag by a few seconds under load. The window is wide enough to catch a burst anyway.

So put the Noxyr role high, below the owner and above everyone else you want covered, and keep Administrator for as few accounts as you can.

A sensible starting configuration

1. Move the Noxyr role up in Server Settings, Roles, just under yours.
2. Set a log channel so every catch is recorded.
3. Enable anti-nuke with the defaults: 3 actions in 30 seconds, timeout for 60 minutes.
4. Add your trusted people with /security trust.
5. Review after a week and move to kick or ban if you want a harder line.

Choosing the punishment

The default is a 60 minute timeout. It stops the damage at once and leaves you time to look, which matters when the account is a real moderator whose phone was hijacked rather than a villain. Kick or ban makes sense for servers where staff turnover is low and any burst of destruction is suspicious.

Setting the action to none turns anti-nuke into an alarm. The bot only logs, so you learn who did what without any punishment. Many owners run it like this for a week to find out which normal admin tasks would trip the limit, and then raise max_actions or trust the right people.

Why it needs to be on before anything happens

Anti-nuke is off by default, and it cannot undo what happened before you enabled it. Switch it on today, even with the log-only action, so the audit trail exists the day you need it.

Test it safely

Use a throwaway server or a test account that is not on the trust list. Delete a few test channels quickly and confirm the log channel gets an entry. If nothing happens, check View Audit Log first, then the bot role position.

Next steps

The settings are in the anti-nuke docs. For the thinking behind them, read the anti-nuke guide, and keep the nuke recovery checklist handy in case it happens anyway. It works best next to anti-spam and the wider security bot. Add Noxyr and turn it on.