Skip to main content

Engage moderation rules

Org admins author regex moderation rules that auto-flag posts and comments across all Engage forums. Posts caught by a rule (and posts manually reported by users) land in a moderation queue for review.

Deep dives for each surface: Moderation rules editor · Flagged content queue · Bulletins authoring

Open moderation rules​

Org → Settings → Moderation Rules.

The moderation rules editor. The moderation rules editor.

Add a rule​

  1. Click Add Rule.
  2. Set:
    • Name — a human-readable label (e.g., "Profanity Filter")
    • Pattern — a regular expression matched against post and comment content. The form validates the regex before allowing save.
    • Reason — optional text shown when the rule fires (helps moderators understand why something was flagged)
  3. Save. The rule is created with is_active = true by default.

When content matches the pattern, a ContentFlag is created with source = auto and status = pending. Manually reported content creates a flag with source = user.

Edit, disable, or delete a rule​

Each rule row has Edit and Delete actions. To temporarily disable a rule without losing it, set is_active to false on the rule (via API today; UI editing covers name / pattern / reason).

Review the flag queue​

Pending flags appear in the org moderation queue. For each flagged post or comment:

  • Approve — content stays visible; flag is resolved.
  • Remove — content is removed; flag status moves to removed.

Each action is recorded against the flag (resolver + timestamp).

What's not built today​

  • The current rule model is a single regex pattern per rule. There are no separate keyword / link / attachment / rate-limit / watchlist rule types, no allow/block/warn action choices per rule, and no per-user watchlist feature.
  • There is no org-level setting for "max posts per hour" or "max @mentions per post" — moderation is post-hoc via regex + flag queue.
  • There is no chapter-level override that lets a chapter tighten the rules independently of org defaults.

Tips​

  • Test your regex carefully. A greedy pattern can flag huge amounts of legitimate content. The form validates that the regex parses, but it doesn't simulate matches.
  • Use a clear reason. Moderators triaging the queue benefit from knowing why something was flagged.
  • Iterate on real data. Pull a sample of pending flags weekly and tune patterns based on false positives.

Last verified against v0.62.1 (2026-05-10).