Moderation
Content automod
Classic automod, built for production. Every rule maps into the same strike and case pipeline as the rest of ModShield — not a separate half-finished system.
- Substring / word / regex lists
- Caps, mentions, emoji, zalgo
- Shared cases & actions
- Safe regex compilation
Why teams turn it on
Blocklists that scale
Substring, whole-word, or regex. Patterns are compiled once and size-limited so a bad expression cannot freeze message handling for the whole server.
More than keywords
Caps ratio, mention spam, emoji spam, zalgo, and duplicate flooding cover the noisy cases keyword lists miss.
One action model
Delete, warn, timeout, or jail — with the same exemptions and cases you already trust on other filters.
Staff-friendly exemptions
Keep automod strict in public channels while mods, bots, and ticket channels stay productive.
How it works
1. Rules you enable
Each automod rule is optional. Turn on only the signals your community actually needs.
2. Match and classify
Messages are checked against blocklists and behavioural rules. Regexes are precompiled for predictable performance.
3. Act through the shared pipeline
Hits become the same deletes, strikes, cases, and jails as link filter or human /warn — one audit trail.
Real situations
Slur / scam word lists
Whole-word mode reduces false positives on substrings while still catching clear abuse terms.
Mention raids
Mention-spam rules stop @everyone-style harassment without disabling mentions for everyone permanently.
Decorative zalgo names/text
Zalgo detection keeps channels readable when attackers paste combining-character walls.
FAQ
Is this Discord AutoMod?
No. It is ModShield’s own engine, shared with Telegram and wired into cases, strikes, and jail.
Can regex take the bot down?
Patterns are compiled once and size-limited. Pathological lists are rejected rather than run unbounded.