How do I protect the server from destructive actions?

Detect and contain dangerous administrative actions while preserving executor and recovery context.

How do I set the safety baseline?

Choose the security log, default response, timeout length, and whether supported damage should be restored. Keep protection off until Yuna has the role position needed to act.

How do I tune each protected action?

Ban, kick, bot add, channel, role, webhook, member-role, server, unban, and mass-mention protections can each follow the baseline or use their own limits.

How do I keep trusted access deliberate?

Trusted members and roles bypass Anti-Nuke. Use this for stable automation or carefully chosen staff, not as a broad fix when a limit feels too strict.

How do I test recovery safely?

Use a controlled non-production action to confirm the executor, affected resource, response, and recovery result are all recorded before relying on the policy during an incident.