Discord global ban list: how to screen raiders and scammers across servers

A Discord global ban list (or shared ban list) lets servers warn each other about accounts that were banned for raids, scams or nukes, so the next server can stop them at the door. Discord itself has no such thing: a ban only applies to the server that issued it, and the same scammer can join your server five minutes later with a clean slate. A shared ban list closes that gap — if it is built carefully, because a list of “bad users” passed between strangers is also a privacy problem and an easy tool for abuse.

This guide explains what a good shared ban list needs, the traps to avoid, and how to set up BotGG Shield: an opt-in network where servers share only counts and categories of bans — never which server banned someone or why — and every server decides for itself what happens when a flagged account joins. It also covers what a ban list can’t catch, and why it belongs next to anti-raid and verification, not instead of them.

What a shared ban list is (and what Discord doesn’t do)

On Discord every server keeps its own ban list. When you ban someone, the ban is stored for your server only; other servers never hear about it. Communities have worked around this in a few ways:

• Staff channels between partner servers — moderators post “heads up, this account scammed our members” with a user ID and a screenshot.
• Ban sync across your own servers — if you own several servers, one ban can be applied in all of them. BotGG’s Server Network does that for servers you run together.
• Shared lists maintained by a group — a spreadsheet, a bot or a website where many servers add the accounts they banned.

All three try to answer one question when someone joins: has this account already caused trouble somewhere else? The answer is valuable for the worst offenders — scam accounts that hop from server to server with fake Nitro links, raid accounts reused across attacks, people who nuke servers after getting staff roles, and ban evaders on their fifth alt.

The risks of a global ban list

A ban list shared with strangers can do real harm if nobody thinks about it:

• False positives. Hacked accounts get banned for scam links they never sent. A list that never forgets punishes the victim twice.
• Grudges. One admin who dislikes someone can put them on a list that dozens of servers trust.
• Privacy. A user ID plus “banned from Server X for reason Y” is personal data. Publishing reasons, server names or screenshots exposes the user and the server that banned them — and can be defamatory if it’s wrong.
• Stale data. A ban from three years ago says little about the person today, and servers unban people all the time.
• Brigading. A hostile group can create servers just to flood a list with bans of people they want excluded.
• Over-reaction. Auto-banning everyone with any mark turns one mistake into a ban from every server that uses the list.

Checklist: what a trustworthy shared ban list needs

How BotGG Shield works

BotGG Shield is built around that checklist. Servers running BotGG can share anonymous ban signals and screen joining members against the signals of other servers. There are two independent switches, both off by default:

• Contribute signals — this server’s qualifying bans are shared with the network.
• Use signals — joining members are scored against other servers’ signals.

You can turn on one, both or neither. Switching either on opens a consent dialog that explains what it means, and the change is recorded in the console’s audit log.

What is shared — and what never is

A signal contains only the banned user’s Discord ID, a category and dates (first seen, last seen, expiry). The categories are raid, spam, scam / phishing, nuke attempt and ban evasion. Generic “rule breaking” is never shared.

Never shared with other servers: which server banned the user, the ban reason text (it is read once for keywords and not stored), messages or attachments, and the moderator who banned. Servers that screen joins only ever see aggregates — per category, how many servers and when first and last seen.

Which bans become signals

Only bans and temporary bans count, and only when the category is clear:

• /ban in BotGG has a shield_category option — pick the category, or none to keep that ban out of the network.
• Bans by anti-raid count as raid, anti-nuke as nuke, AutoMod phishing as scam and other AutoMod bans as spam. Automatic escalation bans (“N warnings”) are never shared.
• Otherwise the reason is checked for keywords in English, Russian and Hebrew (“raid”, “phishing”, “ban evasion” and so on) or tags like #raid and [shield:scam]. You can turn keyword reading off so that only tags count.
• Optionally bans made outside BotGG (in Discord or by other bots) are picked up from the audit log, classified the same way. This needs View Audit Log.

You can also limit which categories your server shares. When you unban someone, that server’s signals about them are deleted. Turning Contribute off removes all of the server’s signals within an hour, and every signal expires 180 days after it was last confirmed.

Trust weighting and abuse protection

Not every server’s ban counts the same. Servers younger than 30 days or with fewer than 50 members don’t count at all; servers that added BotGG less than 7 days ago count half. Size sets the weight — 0.5 for 50–199 members, 0.8 for 200–999, 1 for 1,000 and more — and a paid plan multiplies it by 1.2 (never above 1.2). A server can add at most 30 new signals per hour and 150 per day; the rest are dropped. All servers of the same owner count once, and BotGG staff can block a server that abuses the network, which deletes its signals.

How a joining member is scored

When someone joins a server that uses signals (bots and allow-listed users are skipped), BotGG adds up the signals from other servers inside a window (90 days by default, 7–180). For each independent server the strongest signal counts, as category weight × trust. The weights are raid 1, spam 0.8, scam 1.2, nuke attempt 1.5 and ban evasion 1. The total lands in a score band — low, medium or high (from 1, 2.5 and 4 by default; you can move them).

Example: an account banned for scams in two servers with 1,000+ members and for a raid in a 300-member server scores 1.2 + 1.2 + 0.8 = 3.2 — the medium band.

Set up BotGG Shield step by step

1. Open https://console.botgg.app, pick your server and in the sidebar under Moderation open BotGG Shield. The feature is off by default; its /shield command appears on your server only once it’s enabled.
2. On the Overview tab, read “What is shared — and what never is”, then switch on Use signals, Contribute signals or both, and confirm the dialog.
3. In Settings, choose the Alert channel (a staff-only channel) and the roles to ping.
4. Pick an action for each band: Nothing, Log only, Alert staff, Quarantine (gives a role that hides the server), Verification (sends the member to your Verification review queue), Kick or Ban. The defaults are gentle: low = log only, medium and high = alert staff.
5. Decide on the safety switches: Allow automatic bans must be ticked before any band can ban; Never auto-remove accounts older than turns kicks and bans of older accounts into alerts; DM before a kick / ban sends a neutral message with no server names or reasons.
6. If you contribute, review What this server shares: shared categories, keyword reading and bans made outside BotGG.
7. Save. The Overview tab now shows your contribution — live signals, new in 24 hours, your trust weight, dropped signals and anything that keeps your bans from counting yet (server too new or too small).

Leave automatic bans off until you have watched the alerts for a few weeks. Automatic kicks and bans are capped at 20 per server per hour, never touch the server owner and respect the role hierarchy.

Handling a flagged join

The alert shows the member, account age, score, categories with the number of servers and the first and last dates — nothing about who banned them. Staff decide with buttons: Allow (adds them to your allow list), Release (removes the quarantine role), Kick, Ban (asks for confirmation) or Dismiss. Every flagged join is also listed on the Flagged joins tab, and the Allow list tab lets you add trusted members by ID with a note.

In Discord, staff (Kick, Ban or Moderate Members, Manage Server, or a Shield staff role) can use /shield check user, /shield allow user [note] and /shield unallow user — all replies are private. Flows can react to the trigger “Shield flagged a join” and use the actions “Allow on Shield” and “Check Shield score”.

If you are the one flagged

Anyone can log in with Discord at https://console.botgg.app/shield to see what the network holds about their own account: categories, number of servers, first and last seen and the expiry — no server names. From there they can ask BotGG staff for a review (for example after a hacked account); staff can clear the signals or keep them. A ban in one server is appealed to that server, and the DM includes its appeal link when Ban appeals is on.

What a ban list can’t catch

Be honest with yourself about the limits. A shared ban list only knows accounts that were already banned somewhere, by a server that contributes, for a category it can recognise. It does nothing against:

• Brand-new accounts with no history — exactly what most raid tools use.
• First-time offenders and anyone banned on servers outside the network.
• Bans with no clear category, which are never shared.

That’s why Shield is a filter, not a wall. Put a captcha or rules gate in front of the server with Verification, let anti-raid detect join waves and lock down, and use an account-age join gate for fresh accounts. Shield adds the missing piece: the repeat offender who looks normal but has been banned elsewhere.

FAQ

Can other servers see that my server banned someone?

No. Other servers only see how many servers shared a signal per category and when. Which server banned, the reason text, messages and the moderator are never shared. Only BotGG staff can see the contributing server, for abuse reviews.

Will Shield ban people automatically?

Only if you set it up that way. The defaults only log and alert. A band can ban only after you tick Allow automatic bans, and automatic kicks and bans are capped at 20 per hour per server.

What happens when I unban someone?

Your server’s signals about that user are deleted at once. Signals also expire 180 days after they were last confirmed, and turning Contribute signals off removes all of your server’s signals within an hour.

Can a new server flood the network with fake bans?

Its bans don’t count: servers younger than 30 days or with fewer than 50 members have zero trust, new installs count half, each server is rate-limited, and all servers of one owner count as one.

Does Shield replace anti-raid and verification?

No. A raid usually uses fresh accounts with no ban history, so Shield won’t see them. Use moderation, verification and anti-raid as your base, and Shield on top for known repeat offenders.