Security System

The security system protects your server against nuking attacks, raid waves, and unverified accounts. Navigate to Left sidebar > Security on your server dashboard to configure all settings. The module has three main components: Anti-Nuke (detects and stops destructive actions like mass bans, channel deletions, and role removals), Anti-Raid (detects suspicious join patterns and triggers automated responses), and Verification Gate (requires new members to pass a challenge before accessing the server).

Anti-nuke monitors destructive actions and automatically punishes the perpetrator when limits are exceeded within a time window.

enabled (toggle, default off): Activate the anti-nuke protection.

windowSec (number input, range 5-300, default 10): The time window in seconds. If a user performs more than the allowed number of destructive actions within this window, the configured action is triggered.

whitelistUsers (multi-user input): User IDs that are exempt from anti-nuke detection. Use this for trusted administrators.

whitelistRoles (multi-role picker): Roles that are exempt from anti-nuke detection.

Each limit rule monitors a specific destructive action and triggers a punishment when the limit is exceeded within the time window. Each rule has: enabled (toggle), limit (number, range 1-25), and action (select: none, strip_roles, timeout, kick, or ban).

ban (default: enabled, limit 3, action strip_roles): Triggers when a user bans too many members.

kick (default: enabled, limit 5, action strip_roles): Triggers when a user kicks too many members.

channelDelete (default: enabled, limit 2, action ban): Triggers when a user deletes too many channels. Low default limit because channel deletion is highly destructive.

roleDelete (default: enabled, limit 2, action ban): Triggers when a user deletes too many roles.

channelCreate (default: disabled, limit 6, action strip_roles): Triggers on excessive channel creation.

roleCreate (default: disabled, limit 6, action strip_roles): Triggers on excessive role creation.

antiMention (default: disabled, limit 3, action strip_roles): Triggers when a user mass-mentions across multiple messages.

Instant rules trigger immediately on a single action, no limit threshold needed. Each rule has: enabled (toggle, default off) and action (select: none, strip_roles, timeout, kick, or ban).

botAdd: A bot is added to the server.
prune: A member prune is executed.
dangerousRolePermissionUpdate: A role gains dangerous permissions like Administrator or Ban Members.
dangerousRoleAdd: A role with dangerous permissions is assigned to a member.
vanityChange: The server vanity URL is changed.
serverRename: The server name is changed.
serverIconChange: The server icon is changed.
roleRename: A role is renamed.
channelRename: A channel is renamed.
emojiDelete: A custom emoji is deleted.
emojiRename: A custom emoji is renamed.
emojiAdd: A custom emoji is added.
inviteDelete: An invite link is deleted.

Detection rules that flag suspicious behavior without punishing:

antiInviteLink (toggle, default off): Detect and flag messages containing Discord invite links from non-whitelisted users.

antiGhostPing (toggle, default off): Detect when a user mentions someone and immediately deletes the message.

Anti-raid detects suspicious join waves and responds automatically to protect the server.

enabled (toggle, default off): Activate anti-raid protection.

windowSec (number input, range 5-300, default 30): The detection time window in seconds.

joinThreshold (number input, range 3-200, default 8): The number of joins within the window that triggers a raid alert.

newAccountHours (number input, range 1-720, default 72): Accounts younger than this threshold (in hours) are flagged as new and contribute more to the raid score.

newAccountRatio (number input, range 0-100, default 60): The percentage of new accounts among recent joins that contributes to raid detection.

cooldownSec (number input, range 30-1800, default 120): Minimum time between raid trigger events to prevent alert fatigue.

action (select: none, quarantine, or kick, default 'none'): The action taken on members that join during a detected raid.

alertChannelId (channel picker): The channel where raid alerts are posted.

quarantineRoleId (role picker): The role applied to quarantined members during a raid. This role should restrict channel access.

The scoring system calculates a risk score based on multiple factors:

scoring enabled (toggle, default on): Enable the scoring-based detection system.

scoring threshold (number input, range 20-200, default 75): The risk score that triggers a raid response.

Scoring weights (each range 0-100):
joinPressure (default 40): Weight for the rate of incoming joins.
accountRisk (default 35): Weight for the ratio of new accounts.
volume (default 25): Weight for overall join volume.
botPressure (default 15): Weight for bot account joins.
burstPace (default 10): Weight for burst join patterns.

Presets available: balanced (threshold 75), strict (threshold 65, higher weights), relaxed (threshold 95, lower weights).

Automatically lock down the server during a detected raid:

lockdown enabled (toggle, default off): Enable automatic lockdown on raid detection.

lockdown templateId (select: balanced, strict, or soft): Choose a lockdown preset.

balanced: 5 minutes, 10s slowmode, lock messages and invites, 20 channels.
strict: 10 minutes, 20s slowmode, lock messages and invites, 40 channels.
soft: 3 minutes, 5s slowmode, lock messages only, 12 channels.

durationSec (number input, range 60-3600): How long the lockdown lasts.
slowmodeSec (number input, range 0-21600): Slowmode applied to channels during lockdown.
lockSendMessages (toggle, default on): Prevent members from sending messages.
lockCreateInvites (toggle, default on): Prevent invite link creation.
maxChannels (number input, range 1-100, default 20): Maximum channels affected by lockdown.

Recovery settings:
recovery enabled (toggle, default off): Enable recovery mode after lockdown.
recovery reportChannelId (channel picker): Channel for recovery reports.
recovery pingRoleId (role picker): Role to ping when a raid is detected.

Configure risk level thresholds for tiered responses:

guarded (number input, range 10-180, default 45): The score threshold for guarded status. Below this is normal.

high (number input, range guarded+1 to 220, default 75): The score threshold for high risk.

critical (number input, range high+1 to 300, default 105): The score threshold for critical risk, triggering the most aggressive response.

Require new members to pass a verification challenge:

enabled (toggle, default off): Activate the verification gate.

mode (select: captcha, rules, or hybrid, default 'captcha'): The verification method. captcha shows an image challenge. rules requires typing a specific phrase. hybrid combines both.

applyToAll (toggle, default off): Apply verification to all new members, not just flagged accounts.

accountAgeHours (number input, range 1-720, default 72): Accounts younger than this are automatically required to verify.

quarantineRoleId (role picker): The role assigned to unverified members, restricting channel access.

verifiedRoleId (role picker): The role assigned after successful verification, granting full access.

autoKickUnverified (toggle, default off): Automatically kick members who do not complete verification within the timeout.

challengeTimeoutSec (number input, range 60-86400, default 900): How long members have to complete verification (default 15 minutes).

rulesPhrase (text input, max 120 characters, default 'I AGREE'): The phrase members must type in rules verification mode.

logChannelId (channel picker): Channel for verification event logs.

bypassUsers (multi-user input): User IDs exempt from verification.

bypassRoles (multi-role picker): Roles exempt from verification.

Enable anti-nuke as a baseline protection for all servers — the default limits catch most nuking attempts without false positives. Whitelist your trusted administrators to prevent the anti-nuke from interfering with legitimate administrative actions. Start with the balanced anti-raid preset and adjust thresholds based on your server size. Use the verification gate with captcha mode for the strongest anti-bot protection. Set up the alertChannelId for anti-raid so you can monitor detection accuracy. Enable lockdown only if your server has experienced actual raid attempts, as lockdowns affect all members temporarily.