Security basics

Security Bot Not Working? Setup Troubleshooting and Fixes

When Security Bot does not do what you expect, the cause is usually one of these: its role sits below the accounts it should act on, it lost the Administrator permission, the trigger you are testing is off, or the feature is part of Premium. Below are common symptoms and their fixes. For first-time setup, see the setup guide.

By Sina, Founder, Security BotUpdated (published )12 min read

Quick checks first

SymptomMost likely causeFix
Nobody gets punishedSecurity Bot's role is below the people it should act onDrag its role above every staff role
"I don't have enough permissions"The bot lost AdministratorRestore Administrator on the bot's role
Commands missing in DiscordClient cache or command permissionsRestart Discord; check Server Settings, Integrations
A settings command only sends a linkThe server is on PremiumChange settings on the dashboard
Anti-raid, anti-spam, or AutoMod pages are lockedNo Premium slot on this serverAssign a slot from your account's Premium page, then add the Premium Bot
A deleted channel cannot be restoredRecovery was off when it was deletedTurn Recovery on now so it covers the next one

The dashboard's Alerts page flags the two most common problems on its own: Administrator permission missing and Role hierarchy blocks punishments. Check it first.

Confirm each module works

Before you assume something is broken, test it in a way you can see. Use a real alt account; your own admin account is exempt from most triggers by hierarchy, so it proves nothing.

ModuleTestWhat you should see
Anti-nukeOpen the Anti-Nuke page, or run /settings category:Anti Nuke on a free serverEach trigger with its on or off state, limit, and punishment
VerificationJoin with an alt accountOnly the verification channel visible until you pass
ModerationRun /mute user:@alt time:1mThe alt is timed out for a minute and an entry appears in the moderation log
AutoMod (Premium)From the alt, post a message with more mentions than your Mention Spam limitMessage blocked; an entry in the AutoMod log
Beast ModeOpen the simulator and replay a burst of channel deletesActivation shown once the replayed points reach your score limit

A quiet security log is normal. It only receives anti-nuke warnings and punishments and anti-raid locks, so on a calm day it stays empty. Use the moderation log test above to confirm logging works.

The dashboard at securitybot.gg signs you in with Discord, using only the identify and guilds scopes. It shows one page per module. On free servers, settings can also be changed with slash commands.

On Premium, settings are managed on the dashboard only. Commands such as /whitelist, /setlimit, /enable, /permission, and the verification settings reply with a link to the right dashboard page instead. Moderation commands like /ban, /mute, and /lock work in Discord on every plan.

To use the dashboard or the settings commands you need Administrator and a role above Security Bot's role, or to be the owner. With the bot's role at the top, as recommended, that usually means only the owner can change settings.

TaskCommand on free serversDashboard page
See anti-nuke triggers and limits/settings category:Anti NukeAnti-Nuke
Change a limit or punishment/setlimit action:Banning Members limit:3, /setpunishment action:Banning Members punishment_type:BanAnti-Nuke
Turn a trigger on or off/enable action:Pruning Members, /disable action:Pruning MembersAnti-Nuke
Exempt a role, user, or channel/whitelist add_role role:@Admin category:Roles: Channel-Delete, plus the add_user and add_channel variantsWhitelists
Turn on verification and create its role and channel/verification enable, /verification setup_role, /verification setup_channelVerification
Give a member moderation commands/permission addModeration

Common errors and fixes

Security Bot does not punish someone who crossed a limit

Work through these in order:

  1. The person's highest role is above Security Bot's role, so the bot cannot act on them. Drag the bot's role above every staff role in Server Settings, Roles.
  2. They are the server owner, who is always exempt.
  3. They are whitelisted for that action.
  4. The trigger is off. Prune, admin role assignment, and dangerous permission edits start off.
  5. The count has not reached the limit yet. The punishment fires when the count reaches the limit, and counters reset every 10 minutes. So five bans spread over an hour never reach a limit of 5.

Actions made through Security Bot's own commands are not counted by anti-nuke, because the bot skips its own audit log entries. /ban, /softban, and /kick have a separate cap instead (see below).

The bot says it does not have enough permissions

Security Bot requires Administrator. Without it, anti-nuke ignores every event and dashboard changes fail with "I must have the administrator permission in order to function". Someone may have edited the bot's role after the invite. Restore Administrator on the role or re-invite with the default permissions from securitybot.gg.

Unverified members can see the whole server

Role order does not decide channel access. If any role a member holds allows View Channel on a channel, they see it, wherever that role sits in the list. Verification works by giving the Unverified role a View Channel deny on every channel, of every type.

Check three things:

  • No role that unverified members hold (including @everyone) has an explicit View Channel allow on the leaking channel.
  • Channels you created after setup got the deny. /verification update_role on free servers, or Update role on the Verification page, adds it to any channel that does not have one yet.
  • The channel is not out of sync with its category, since a channel's own overrides decide its access.

Security Bot's role must also sit above the Unverified role so it can remove it.

Slash commands do not appear

The applications.commands scope is included with the bot invite by default, so this is rarely an invite problem. Usually the Discord client has a stale command list: press Ctrl+R on desktop, or force-close and reopen the app on mobile. If commands are still missing for some members, check Server Settings, Integrations, Security Bot, where commands can be limited to certain roles or channels.

A moderator gets "You failed to pass the requirements to use this command"

To run a moderation command, a member needs the matching Discord permission and a role above Security Bot's role, or to be the owner, or a command permission granted to them on the Moderation page. With the bot's role at the top, most moderators need the Moderation page grant.

/ban says "You have reached the ban limit, try again later"

While the Anti Ban and Anti Kick triggers are on, /ban, /softban, and /kick (and /banid on Premium) are capped per moderator. The cap is your anti-nuke ban and kick limits, and the counters reset every 10 minutes. Nothing else happens at the cap: no punishment, no owner alert.

If your team needs more during cleanups, raise the ban or kick limit on the Anti-Nuke page. Whitelisting does not raise the cap. The owner and Administrators ranked above the bot are not capped.

The member logged into securitybot.gg with a different Discord account than the one that joined, or they are not yet in the server. They should log out of the website, log in with the joining account, and pick the server from the list.

Anti-nuke punished a staff member doing normal work

Their count for that action reached the limit before the 10-minute reset, or the role needs an exemption for that one action. Check the security log for which trigger fired.

If one role or bot does that action in bulk as part of its job, whitelist it for that action only. Raise the limit on the Anti-Nuke page only if your whole team's normal volume is above it. Give the roles back if the punishment was clear roles. Do not switch the trigger off, because that removes the protection for everyone.

An AutoMod preset says it conflicts with another

Discord allows one rule per trigger type for spam, mention spam, flagged words, and member profile, plus six keyword rules. When you turn on a preset that needs a slot another one uses, the dashboard switches the other preset off and tells you. If you created rules by hand in Server Settings, AutoMod before adding the bot, delete the duplicates there and re-enable the preset.

Anti-spam settings will not save

The bot rejects an allowed-messages value below 5 ("Allowed messages must be at least 5.") and a duration below 8 seconds ("Allowed messages delay must be at least 8 seconds."). Remember that anti-spam counts every message in a channel, not per member. So the default of 20 messages in 8 seconds is a better starting point than a very low limit.

Anti-spam keeps locking a busy channel

Because the count covers everyone in the channel, lively conversation can reach the limit. Raise the limit, or add the channel to the Channels: Spam whitelist.

If locked channels never reopen, the lock duration is 0, which means the lock stays until someone runs /unlock. Set a duration.

Premium features are locked

Premium is assigned per server. Open Premium under your account on securitybot.gg, choose which servers use your slots, and save.

Premium modules run on a separate Discord app, the Premium Bot. After assigning the slot, invite that bot to the server and reload the server dashboard. It leaves servers that have no slot. Settings made with the free bot do not carry over, so set the modules up again on the dashboard.

An existing Patreon membership with your Discord account linked can take a few minutes to sync. If you cannot find the Premium Bot's invite, ask in the support server.

A deleted channel or role cannot be restored

Recovery (Premium) is off until you turn it on, with the switch on the Recovery page or /recovery enable. Anything deleted before you turned it on cannot be restored. Turn it on now.

From then on it takes automatic backups of the server layout (every 24 hours by default). It also keeps every deleted role, channel, category, and emoji under Recently deleted for 90 days.

Restore from the Recovery page, or with commands: /recovery restore for a backup, and /recovery role, /recovery text, /recovery voice, or /recovery category for deleted items by ID.

Beast Mode never activates

Every action score starts at 0 and the default score limit is 40, so out of the box nothing adds points. Set action scores on the Beast Mode page, then replay an attack in the simulator to check that it activates where you expect.

Something is missing from the logs

Free servers get the moderation and security log channels. The audit and AutoMod log channels are part of Premium.

Some things are never logged. /clear and /raid ban are not written to the moderation log. Verification passes, Anti Invite Link deletions, anti-spam locks, anti-raid bans, and Beast Mode activation are not logged either.

The security log receives anti-nuke warnings and punishments and anti-raid locks. Anti-raid timeouts and the channels anti-raid locks appear in the moderation log, with Security Bot as the moderator.

The bot went offline

Your Discord permissions are unchanged and anti-nuke resumes when it reconnects. This is why the native side of the setup matters: staff without Ban Members cannot mass-ban during an outage. The trust page covers what the bot stores and why it requires Administrator.

If you are being attacked right now

Do not read the rest of this page first.

  1. Invite the bot from securitybot.gg.
  2. Drag its role above the attacking account.
  3. Run /lockall to freeze every text channel.
  4. Pause invites in Server Settings, Invites.

Anti-nuke acts on the next actions it sees once it has Administrator and the higher role. It cannot undo what happened before it arrived.

Once the attack stops, read the security log to see who was punished. Unban real members caught in a ban wave, then come back to the checks above.

Tuning after the first week

  • A staff member tripped a limit doing normal work. Whitelist their role for that action only. Raise the limit only if the whole team's normal volume is above it.
  • Anti-raid locked channels on a normal day. The score limit is too low for your usual join pattern. Raise it, or lower the normal-join score so clean joins add less.
  • Real members failed verification. Lengthen the captcha timeout (30 seconds to 5 minutes), shorten the code (4 to 8 characters), or switch to one-click verification.
  • AutoMod blocked normal messages. Add allow terms to the preset, or exempt the channel.
  • Nothing in the security log at all. That is normal on a quiet week. Run a short /mute on an alt to confirm the moderation log works.

Month one: a maintenance routine

After the first weeks the settings should be stable, and the job becomes keeping them true. This routine catches drift before it becomes an incident.

  • Weekly: read the security log and the AutoMod log. Add allow terms for false positives. If a limit fired on staff, whitelist that role for the action, or raise the limit if the whole team's normal work reaches it.
  • Monthly: review the Whitelists page, or on free servers run /whitelist view_roles, /whitelist view_users, and /whitelist view_channels with a category: for each list, and remove entries for staff who left. Open Server Settings, Integrations and check that no bot gained Administrator. Confirm the bot's role is still at the top.
  • After staff changes: a new moderator gets command access on the Moderation page, not Ban Members on their role. A departing one loses the role the same day, because a whitelisted role keeps its exemption for whoever holds it.
  • On Premium: confirm Recovery is on and that recent automatic backups appear on the Recovery page. It only has what it saved while it was on.
  • Quarterly: run the interactive security audit and compare its score with the one on the dashboard overview. Fix whatever gap both of them flag.

What to do next

  1. Read the feature pages for the modules you enabled; each has a command reference and setup steps.
  2. Bookmark the security log channel and check it daily for the first two weeks.
  3. Join the support server from the footer if something in this list does not fix your problem, or search the documentation.

Do this with Security Bot

Discord Anti Nuke Bot: Stop nukes with per-user action limits, automatic punishments, owner alerts, and a security log.

Security Bot is a Discord anti nuke bot that watches bans, kicks, channel and role changes, bot adds, prunes, and dangerous permission edits. When an account hits a limit, it is kicked, banned, or stripped of its roles. The server owner gets a DM.

Frequently asked questions

Why is Security Bot not punishing anyone?

Usually role hierarchy: the bot's role is below the staff roles it should act on. Move it to the top. Also check that the trigger is on, that the person is not the owner or whitelisted, and that their count reached the limit before the counters reset (every 10 minutes).

Why don't Security Bot's slash commands show up?

The applications.commands scope is included with the invite by default, so it is usually a stale client. Press Ctrl+R on desktop or restart the mobile app, then check Server Settings, Integrations, Security Bot for command restrictions.

Why does a settings command only send me a dashboard link?

On Premium, settings are managed on the dashboard only, so commands like /whitelist and /setlimit reply with a link. Free servers can use the commands directly. Moderation commands work in Discord on every plan.

Why can't I restore a deleted channel?

Recovery (Premium) is off until you turn it on, and it only has the backups and deleted items it saved while it was on. Turn it on with the switch on the Recovery page or /recovery enable so the next deletion can be restored.

How do I test that Security Bot works?

Use an alt account: join to test verification, run a one-minute /mute to see the moderation log, and on Premium post more mentions than your Mention Spam limit to test AutoMod. Use the Beast Mode simulator for anti-nuke scenarios.

Where do I get help with Security Bot?

The documentation at docs.securitybot.gg and the support server linked in the site footer.

How this guide was made: drafted with AI assistance, then checked against Security Bot's source code and Discord's official help articles. Spotted something out of date? Tell us in the support server.