Raid mode

Raid mode is an authorized, time-bounded incident posture. An operator records a reviewed manual or join-burst trigger, owner, duration, and severity. For automatic join-burst activation, the reviewed owner and severity are saved in the revisioned raid policy before the switch can be enabled; Pup does not invent an owner when a burst occurs. Pup snapshots the exact prior join-admission policy revision and provider state before applying the configured tighter posture for join admission, verification, messaging, invites, roles, and notifications. All policy values are bounded by the dashboard and API provider ceilings. Recovery is staged and revision-aware: every provider surface is compared with the exact value Pup applied, then restored only when it still matches. Role changes are paused and member-facing notifications are suppressed while their respective tightened postures are active; staff alerts can still be delivered. A manual change made during an incident is retained and reported as preserved rather than overwritten. Open Protection overview → Raid mode to configure the policy, preview activation, and review recovery. Activation and recovery require current server configuration access.

​

Configure and start a raid response

Review Saved raid response before activation. Activation settings controls manual-only or automatic join-burst activation, the accountable moderator, burst threshold/window and duration. Response settings controls Discord verification, channel slowmode, new invites, Pup’s automatic anti-nuke roles and join-guard notifications. Save the configuration before starting an incident. The current implementation raises Discord verification to at least High, applies at least 10 seconds of slowmode when selected, and denies new invites for @everyone in server channels. Existing invites and role-specific overrides are not revoked. Both staff-alert choices suppress member-facing join-guard notices; neither creates a new audit channel. A legacy review-only messaging setting also applies slowmode; it does not hold messages for approval. Choose a responsible moderator, confirm the saved response and activate. Manual incidents use High severity. Activation in progress means Discord work is queued or unfinished; Raid mode active confirms that the provider changes were recorded. Refresh status loads the current state without discarding configuration drafts.

​

End and recover

End raid and restore settings starts recovery; expiry also queues recovery. Restoration in progress remains visible until recovery completes, including after the scheduled end time. Settings cannot be edited in this dashboard while the incident is open. Recovery uses the incident’s original join policy and Discord snapshot, even if the saved raid policy changed later. If no join policy existed before the incident, recovery removes the temporary policy. Partial Discord writes can be recovered before activation completes. Each channel is compared independently; later manual changes are preserved, and a retry does not reapply a completed restoration. Only Restoration complete confirms the recovery operation finished.