Community Safety Desk

Server structure

Overview

On-chain communities are targeted constantly, and most servers treat safety as a pinned message. This layout makes it the structure. #official-links-only is read-only and lists every genuine address, domain and account the project uses, with the instruction to check here before clicking anything, ever. #staff-will-never is the companion: a short, blunt list of things nobody legitimate will ever do, written so that a member under pressure can recognise the pattern. #verify-a-link is where anyone can paste something and ask before interacting, which is the single most useful channel in a server like this and works precisely because asking is normalised rather than embarrassing. Reports covers what has already happened: #report-a-scam, #impersonation-log, a read-only running list of fake accounts and lookalike domains, and #known-bad, the accumulated list. Rescue is the channel most projects lack entirely: #i-think-i-was-compromised gives the immediate steps in order, calmly, at the moment somebody is panicking and most likely to make it worse. Education keeps #how-approvals-work, #wallet-hygiene and #social-engineering-patterns, because the recurring failures are behavioural rather than technical. A hidden desk area handles takedowns and coordination with moderators. Nothing here discusses price or offers financial advice.

When to use it

Communities in this space lose members to the same handful of attacks: a lookalike domain, a fake support account, an approval nobody understood. Most servers respond with one pinned message and a moderation team that deletes links after the fact. This layout publishes the verified list, normalises asking before clicking, keeps a running impersonation log, and gives someone who has just been compromised clear steps instead of panic.

What makes it different

• One read-only channel lists every genuine link, so verification has a single source.
• Asking whether a link is safe is normalised in its own channel rather than embarrassing.
• A running impersonation log, so a fake account is recognised the second time by everyone.
• Clear first steps for someone who has just been compromised, written before the panic.

Recommended Vetox setup

• Auto-Mod

  Run Auto-Mod link filtering everywhere, so a drainer link posted at 3am is removed before a moderator wakes up.

• Anti-Raid

  Enable Anti-Raid, because coordinated joins ahead of an impersonation wave are the standard opening move here.

• Tickets

  Point Tickets at #i-think-i-was-compromised so a distressed member gets a private thread rather than a public audience.

• Logs

  Enable Logs so role, nickname and channel changes are visible, which is how a compromised moderator account is caught.