Role and Permission Design
Use separate roles for Moderators and Admins. Enable only the permissions they need (timeouts, kick, ban, manage messages). Avoid Administrator for routine moderation.
AutoMod Baseline Rules
- Block spam: mention spam, repeated messages, excessive emojis
- Filter profanity and slurs with keyword lists and allowlists
- Prevent malicious links: block IP grabbers and shorteners
- Require media in media-only channels
Actions and Escalation
- Timeout on first violation; increase duration on repeats
- Delete message, notify user with reason, log to a mod channel
- Escalate to kick/ban for raiding or coordinated abuse
Logging and Evidence
Audit logs and a private #mod-logs channel are essential. Store message snapshots and reasons for actions to resolve disputes.
Appeals and Transparency
Create a ticket channel or form for appeals. Publish a short server ruleset and a moderation policy so users know expectations.
AutoMod Presets and Custom Rules
AutoMod supports both Commonly Flagged Words (Discord-maintained lists) and Custom Keyword Filters (up to three custom lists). Presets include categories like Insults & Slurs, Sexual Content, and Severe Profanity. Use custom lists for project‑specific terms, scams, and community jargon misuse. Wildcards (*) match partials, but tune carefully to avoid false positives.
Mention and Content Spam
- Mention spam: Configure a maximum number of unique user/role mentions per message (up to 50). Recommended response: Block message to prevent notifications.
- Spam content: Enable Discord’s model‑driven spam filter to block widely reported scams and unsolicited content across text channels, threads, text‑in‑voice, and forums.
Responses and Logging
- Block message: Prevents the message from posting and shows the user an ephemeral notice.
- Send alert: Forwards details to a mod log channel with quick actions.
- Timeout: Temporarily restricts the offender (no messages, reactions, voice join).
Exemptions
Define roles, channels, and specific allowed words that bypass filters. Users with Manage Server or Administrator, as well as bots and webhooks, are exempt by default.
Operational Tips
- Use alerts for subjective content requiring human review; block for slurs and severe profanity.
- During raids, raise mention limits sensitivity, temporarily pause invites, and increase verification level.
- Document a moderation policy, appeal process, and keep evidence logs for transparency.