Yorigard

Server security, without the noise — and without the illusions.

Yorigard protects Discord servers from the attacks that really hurt: destruction by a compromised account, coordinated raids, spam. With one simple rule — never let you believe a gate is closed when it isn't.

What it stops

• Anti-nuke — mass deletion or creation of channels and roles, ban sprees, granting of critical permissions. The costliest threat, because it is irreversible.
• Anti-raid — the coordinated arrival of dozens of accounts. The response is a server lockdown, not a witch hunt.
• Anti-spam — message floods, mention spam, repeats. A nuisance without destruction, met with a temporary, reversible timeout.

Sanctions you check before anyone feels them

Every detector can warn, timeout, kick or ban — but nothing is left to chance:

• Simulation mode, on by default. Yorigard notes what it would have done, without touching anyone. You read, you adjust, then you turn it on.
• Immunity threshold — your admins and trusted members never trip the alarms.
• Circuit breaker — a burst of sanctions within seconds betrays an over-aggressive setting: the system stops on its own rather than punishing in cascade.

Lockdown, when a wave arrives

Faced with a raid, Yorigard raises the server's verification level for as long as the wave lasts. No sanctions, no one to single out: the gate closes, then reopens.

A gate on arrival

An optional verification filters newcomers:

• Anti-bot check (Cloudflare Turnstile) — stops automated raids.
• Minimum account age — the only honest signal against throwaway accounts.
• Custom link — a short address to share in a channel, for members who block direct messages.
• Report to a channel — see every attempt as it happens.
• A member who leaves the server goes through the gate again if they return.

Everything is visible, everything is in reach

• A clear, fully bilingual (French / English) web dashboard: settings, logs, verifications.
• A security log and an audit log — who did what, with the old and new value, kept even if the bot leaves the server.
• A /dashboard command to find the link at any time.

Built for trust

Yorigard never asks for the Administrator permission. It requests exactly what it needs, and nothing more: should its access ever leak, the damage stays bounded. Security that doesn't become a risk itself.

Add it, let it watch for a week in simulation, then enable what you've reviewed. Protection should never punish on the strength of settings no one has read.