Automation & insights
AutoMod suggestions
Suggestions surface risk patterns without auto-punishment. Humans stay in control while the system detects trends.
Commands
,suggest view
,suggest acknowledge <id>
,suggest action <id> <text>
Suggestion types
- Repeat warnings within 7 days
- Escalation to kick patterns
- Escalation to ban patterns
- Staff overuse alerts
- Behavioral anomalies
Schema highlights
Core fields stored in the automod_suggestions table.
guild_id
Server scope for the suggestion.
target_user_id
VRChat user flagged by the system.
suggestion_type
Pattern category name.
confidence_score
0-100 confidence rating.
details
JSON payload describing the pattern.
resolved
Whether staff resolved the suggestion.
acknowledged_by
Staff member who reviewed it.
action_taken
Logged response or notes.
created_at
Timestamp when generated.
Walkthrough
- Run ,suggest view to list new recommendations.
- Use ,suggest acknowledge to mark a suggestion reviewed.
- Log follow-up actions with ,suggest action.
- Track confidence scores before taking action.