Security basics

Discord Server Security Checklist: 20 Settings to Check

Securing a Discord server comes down to twenty settings across five areas: accounts (2FA), permissions (least privilege), the door (verification level, captcha, raid protection), inside (anti-nuke, spam filters, logging), and recovery (a way to restore what gets deleted). Most take a minute. This page is the full list. For a quick score, use the interactive security audit.

By Sina, Founder, Security BotUpdated (published )11 min read

Accounts

  1. Owner has 2FA. User Settings, My Account, Password and Authentication. Discord recommends a security key or passkey. An authenticator app also works. The owner account is the one thing no bot can protect.
  2. 2FA is required for moderation. Turn on the server's 2FA requirement for moderation in Server Settings. Staff without 2FA lose their native moderation permissions until they enable it. It stops attackers who only have a password, not a hijacked session such as a stolen token or an approved QR-code login.
  3. Staff use unique passwords. Token stealers and reused passwords are common ways staff accounts are lost.

Permissions

  1. Only the owner has Administrator. Server Settings, Roles. Everyone else gets specific permissions.
  2. Manage Server, Manage Roles, and Manage Channels are limited to one trusted admin. These are the nuke permissions. See dangerous permissions explained.
  3. Moderators do not hold Ban Members or Kick Members natively. They moderate through Security Bot with per-user command permissions. Its ban and kick commands are capped per moderator while Anti Ban and Anti Kick are on (the default).
  4. Bots have the minimum permissions they use. Server Settings, Integrations. Re-invite any bot that was given Administrator for convenience.
  5. Security Bot's role is above every staff role. Anti-nuke can only act on accounts below it.

The door

  1. Verification level is Medium or High. Server Settings, Safety Setup. High means a member must have been in the server for 10 minutes (plus a verified email and an account older than 5 minutes) before posting. A verified phone number skips these requirements, so the level alone does not stop phone-verified raid accounts.
  2. Verification is on. New members get an Unverified role and must pass a check, such as Adaptive Verification, an image captcha, or the web check. That raises the cost of each automated account. Free in Security Bot. See how to set up verification.
  3. Raid protection is on. Discord's own Raid Protection (Server Settings, Safety Setup, Raid Protection and CAPTCHA) detects join raids, sends alerts, and challenges new joiners afterwards. On Premium, Security Bot's anti-raid adds join scoring with your own weights. It times out or bans the accounts that joined in the raid window and locks down chosen channels.
  4. Anti Bot Add is on. Anyone below Security Bot who adds a bot application is punished and the bot is removed, unless they are whitelisted for it. Free, and on by default.
  5. Server Discovery and public listings are deliberate. If you do not need to be discoverable, turn Discovery off under Community.

Inside

  1. Anti-nuke limits and punishments are reviewed. The defaults are 5 bans, 5 kicks, 3 channel or role deletes, and 8 channel or role creates per user per 10 minutes, with kick as the punishment. Turn on the free instant triggers that start off: prune, admin role assignment, and dangerous permission edits.
  2. Beast Mode is enabled with action scores set. Every action score starts at 0, and the default score limit is 40. Give points to the actions you care about, or it will never activate.
  3. AutoMod rules are on. Mention spam, spam content, and commonly flagged words. Set them up in Server Settings, AutoMod, or from Security Bot's presets on Premium. The presets also add the Threat Link Guard scam filter.
  4. Anti Invite Link is on so invite spam in regular text channels is deleted automatically (threads, forum posts, announcement channels, and voice-channel chat are not checked). It deletes every invite, including your own, so whitelist the channels where invites belong.
  5. Log channels are assigned and visible to the right staff: moderation and security logs on every plan, plus audit and AutoMod log channels on Premium.

Recovery

  1. A recovery path exists before you need it. On Premium, turn Recovery on from the dashboard or with /recovery enable. It stays off until you do. A Server Template is a limited free fallback.
  2. An incident plan is written. Who runs /lockall, who pauses invites, who restores, and where the log is. One pinned message in the staff channel is enough.

Once Recovery is on, Security Bot backs up roles, categories, channels with their permission overrides, emojis, and who had which role (every 24 hours by default). It keeps each deleted role, channel, category, or emoji in a Recently deleted list for 90 days. Restores show a preview first.

Messages cannot be restored, and nothing from before you turned Recovery on is saved. A Server Template snapshots channels, roles, and permissions, but it must be synced by hand and can only create a new server.

If you only have fifteen minutes

Owner 2FA, the 2FA requirement for moderation, owner-only Administrator, limiting the Manage permissions, the verification level, Anti Bot Add, the anti-nuke review, and the log channels are free and usually take under fifteen minutes combined. Start with those even if you do nothing else.

Score yourself

CheckedStatus
18 to 20Hardened. Review quarterly and after staff changes.
12 to 17Reasonable. Close the gaps in the door and inside sections first.
Under 12Exposed. Start with 2FA, Administrator cleanup, verification, and anti-nuke limits.

For a quick score from a few answers, use the interactive security audit. Or invite Security Bot and open the dashboard. The overview page shows a security score for your server, and the alerts page lists the specific gaps it found.

What Discord provides and what needs a bot

Half of this list is native. Discord ships 2FA, roles and channel overrides, verification levels, Safety Setup with Raid Protection and CAPTCHA, AutoMod, the suspicious-link filter in Community servers, and an audit log that keeps entries for 45 days.

What it does not ship is anything that counts actions per user. Nothing native stops an admin deleting thirty channels in a row, punishes or locks down a raid with your own scoring, or restores a deleted role.

NeedDiscord nativeSecurity Bot
Account security2FA and the 2FA requirement for moderationNothing; no bot can do this for you
Least privilegeRoles and channel overridesPer-user moderation commands so staff never need Ban Members
Gate on new joinsVerification level, with wait times and phone or email checksAdaptive, web check, image captcha, one-click, passcode, or allowlist verification with an Unverified role (free)
Raid detectionRaid Protection: join-raid detection, Activity Alerts, and CAPTCHA for new joiners after a raidJoin scoring with your own weights, punishment, and lockdown (Premium)
Nuke preventionThe 2FA requirement and bulk-ban prompts add friction, but nothing caps actions per userPer-user limits with kick, ban, or clear roles (free)
Content filteringAutoMod rules, configured by hand, and the suspicious-link filterAutoMod presets that create the rules, plus bot timeouts, kicks, or bans and a scam domain feed (Premium)
Audit trailAudit log, 45 days, no alertsModeration and security log channels (free); audit and AutoMod log channels (Premium)
RecoveryServer Templates, synced by hand, that can only create a new serverRecovery, once turned on, takes automatic backups and keeps recently deleted items, with a preview before each restore (Premium)

Small community under 500 members

Do the accounts and permissions sections completely. They cost nothing, and a small server lives or dies by its owner account. On the door, set the verification level to Medium and turn on Adaptive Verification on Relaxed, or one-click.

Inside, a ban and kick limit of 2 or 3 with clear roles as the punishment is a reasonable start. Anti-raid scoring and Recovery can wait until a public invite is floating around or the layout would take more than an hour to rebuild.

Public gaming server

Public invites bring a steady flow of unknown accounts. Set the verification level to High, use Adaptive Verification on Balanced, and keep Discord's Raid Protection on. On Premium, turn on anti-raid with lockdown.

Enable Beast Mode with action scores set. Turn on the native AutoMod rules plus Anti Invite Link, since invite spam is common on public servers. Give moderators moderation commands through the bot instead of Ban Members.

Crypto, NFT, or marketplace server

Your members are the target and scam links are the weapon. Use Adaptive on Strict, or the web check, with VPN and alt blocking. On Premium, use:

  • Anti-raid with a score limit below the default 200, set from your normal join rate. Every clean join adds 2 points, and everyone who joined inside the window is punished when the limit is reached.
  • Threat Link Guard for scam and phishing domains.
  • Recovery, turned on.

Use the ban punishment on the channel and role delete triggers. Keep whitelists short and never global, and keep Allow whitelist bypass off on Beast Mode. This is the profile where the Premium modules matter most.

Doing the list in order on a new server

If you are starting from an empty server, the order matters less than finishing, but this sequence avoids redoing work.

  1. Enable 2FA on the owner account and require it for moderation before you invite anyone else.
  2. Create roles with specific permissions before assigning anyone. It is much easier than removing Administrator from people who got used to it.
  3. Invite Security Bot using the setup guide and drag its role to the top immediately. Everything below it is protected. Anything above it is not.
  4. Turn on verification and let the bot create the Unverified role and channel, then create your normal channels. On free servers, /verification update_role adds the Unverified overrides to channels of any type you create later. On Premium, use the Verification page.
  5. Set log channels, review anti-nuke limits on the Anti-Nuke page, and whitelist your admin role for channel and role actions only.
  6. Turn on AutoMod rules and Anti Invite Link last, once you know which channels are for what, so the exemptions are right the first time.

Common mistakes when working through the list

  • Ticking the permissions item without checking bots. A bot invited with Administrator two years ago is an admin account nobody logs into. Server Settings, Integrations lists every one.
  • Setting the verification level to Highest and calling it done. Highest requires a verified phone, which shuts out real members without one and does not stop phone-verified raid accounts. A lower level plus captcha verification adds a check without the phone requirement.
  • Leaving Security Bot's role in the middle of the list. Every role above it is exempt from every trigger. This is a common reason anti-nuke appears not to work.
  • Global whitelists for the staff role. One phished moderator and anti-nuke is off for them. Whitelist per action. See how to whitelist safely.
  • Log channels visible to everyone. The security log shows who was punished and why. Keep it in a staff category.
  • No test after setup. Join with an alt, run a short mute, replay a nuke in the Beast Mode simulator. A setting you have not tested is a setting you assume works.

A weekly and monthly routine

Security settings decay. Staff change, bots get added, and someone turns a trigger off during an event and forgets. A short routine keeps the twenty items true after you tick them.

WhenDo
WeeklyRead the security log. A trigger that fired on staff means a limit is too low or a role needs a per-action whitelist. An anti-raid lock on a normal day means the raid score limit is too low.
WeeklySkim the AutoMod log or alert channel for false positives and add allow terms to the flagged words rule.
MonthlyOpen the Whitelists page or run the whitelist view commands, and remove entries for staff who left and bots you uninstalled.
MonthlyOpen Server Settings, Integrations and Roles. Remove Administrator from anything that gained it and confirm the bot's role is still on top.
QuarterlyRedo the full list, run the interactive security audit again, and update the pinned incident plan with current staff names.
After any incidentRead the log timeline, unban members caught in a ban wave, restore from a backup or the Recently deleted list if Recovery was on, and tighten whichever limit the attacker got past.

Security Bot's logs are ordinary messages in your own Discord channels, so they stay until someone deletes them. Keep the channels private, and use Discord search on them when you need to look back at an incident. The logging page lists what each channel receives.

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

How do I add security to a Discord server?

Enable 2FA, limit Administrator and Manage permissions, set the verification level, add captcha verification, turn on anti-nuke limits and AutoMod, assign log channels, and set up a recovery path. Security Bot covers the bot-side items from one dashboard.

What is the most important Discord security setting?

Requiring 2FA for moderation and keeping Administrator with the owner only. A nuke needs a trusted account's permissions, so protecting those accounts matters most.

Is there a free way to secure a Discord server?

Yes. Every native setting on this list is free, and Security Bot's anti-nuke, verification, moderation commands, invite filter, Beast Mode, and moderation and security logs are free.

How often should I review server security?

Quarterly, after any staff change, and after any incident. Whitelists and bot permissions drift the most.

Does Discord have a built-in security checklist?

Safety Setup groups some settings, but it does not cover bots, anti-nuke limits, or recovery. This list does.

Do I need all twenty items on a small server?

Do the account and permission items regardless of size. They are free, and they are where many compromises start. Anti-raid scoring and Recovery matter once you have public invites or a layout that would take hours to rebuild.

How does the interactive security audit score my server?

It asks for server size, whether invites are public, staff count, verification type, logging level, whether Recovery is on, and anti-raid. Then it returns a score from 15 to 100 with the specific gaps and the fix for each.

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.