How moderation works

Moderation runs on a simple, trustworthy loop: the employee watches, judges, and flags, then a human confirms the action.

Every message in a watched channel is read for meaning, not just for banned words. When something appears to break a rule, the employee snapshots the evidence and posts a flag with the reason and the rule it matched.

From the flag, a moderator can warn, delete, time out, or dismiss with one click. The action is recorded permanently, and the incident is marked resolved. The employee never takes a consequential action on its own.

This is placeholder documentation for the foundation build.