Nova Guardian
Nova Guardian is a passive community-protection shield. Once an admin turns it on, Nova silently screens every message in your Discord server or Telegram group for scam intent and social engineering — not just links — and takes graduated, configurable action.
The hard part of community safety isn't banning links (servers already do that) — it's the text-based lures that carry no link at all. Guardian reads intent with an LLM, so it catches paraphrases, typos, leetspeak and letter-spacing tricks that keyword blocklists miss.
Guardian is admin/moderator-only and group-only — it can't be set up in a DM, and only a server/group admin or moderator can enable or change it. Turning it on protects the whole server/group, not just one channel.
What It Catches
🎣 Fake hiring / recruiting
"We're onboarding mods/testers/ambassadors — DM me", fake remote web3 jobs, pay-per-task lures
🪂 Fake airdrops / claims
"Claim your airdrop", "verify your eligible wallet", "check eligibility", fake free-mints
🔑 Seed / wallet harvesting
Asking for or "buying old/inactive wallets", "paying for seed phrases", "import/sync/restore your wallet"
🎭 Team / admin impersonation
"Official support here, DM me to resolve", or a name that mimics a real admin
📨 DM-bait / off-platform lures
"Let's continue in DM", "message me on WhatsApp/Telegram/Signal"
💸 Send-to-receive / advance-fee
"Send 0.1 get 0.2 back", "double your crypto", "first 50 only", guaranteed-return giveaways
🔗 Malicious / phishing links
Lookalike domains, wallet-drainer sites, fake claim/mint pages, hidden URL shorteners
Guardian judges intent, not keywords. Warning others about a scam, asking how airdrops work, or legitimate OTC trading talk are recognized as clean — and you can allowlist phrases that are normal for your community.
How It Decides — the action ladder
Every message gets a verdict, and each tier has its own action:
Clean
Nothing.
Suspicious
Silently posts a mod-review card to your review channel — author, age, reason, message link. No public action.
Scam (high-confidence)
Posts the review card and (only if you've turned on enforcement) auto-deletes the message and mutes/timeouts the author.
Enforcement is OFF by default. When you first enable Guardian it only flags — it never deletes or mutes anyone until you explicitly turn on auto-delete / auto-mute. Start in flag-only mode, watch the review channel, then enable enforcement once you trust the calls.
Sensitivity modes
Lenient
Only the most blatant scams are flagged — fewest false positives.
Balanced (default)
Sensible middle ground.
Strict
Catches more, flags more for review — a few more false positives for mods to dismiss.
Trusted, established members are held to a lighter standard than brand-new accounts, so a long-time member and a fresh account posting the same line aren't treated the same.
Set It Up — just talk to Nova
There's no wizard — you configure Guardian conversationally. In your server/group (as an admin/mod), @mention Nova or reply to it:
Turn it on:
"@Nova turn on Guardian / scam protection for this server"
Nova enables flag-only screening for the whole server/group and posts review cards to the channel where you enabled it.
Point it at a mod channel (optional):
"send Guardian flags to #mod-review" (Discord) — defaults to the enable channel otherwise.
Tell it who the real admins are (so it can catch impersonators):
"the real admins are @alice and @bob"
Add allowlist phrases that are normal for your community:
"allowlist 'buying wallets' — this is an OTC channel"
Turn on enforcement when ready:
"enable auto-delete and auto-mute for scams" · "mute scammers for 2 hours"
Tune sensitivity / check status anytime:
"make Guardian stricter" · "is Guardian on?" · "how many scams did Guardian catch this week?"
"turn off Guardian"
Bot permissions for enforcement. To auto-delete and mute, the bot needs the right permissions — on Discord: Manage Messages + Moderate Members; on Telegram: make the bot an admin with the restrict members right. If a permission is missing, Guardian falls back to flag-only and tells you exactly what to grant.
Configuration Options
Mode
strict / balanced / lenient detection sensitivity
Mod-review channel
Where flag cards are posted (defaults to where you enabled it)
Auto-delete
Auto-remove high-confidence scam messages (default off)
Auto-mute
Timeout/mute the author of a high-confidence scam (default off)
Mute duration
How long an auto-mute lasts (default 60 minutes)
Allowlist
Phrases that must never be flagged (normal-for-your-community talk)
Known admins
The real team identities, so Guardian can flag impersonators
Thresholds
Advanced numeric overrides for the suspicious / scam cut-offs
Weekly Recap
Every Monday at 12:00 UTC, Guardian posts a short recap to your mod-review channel: how many messages it flagged that week, how many were high-confidence scams, what actions it took, and the top patterns it saw. (It stays quiet on weeks with nothing to report.)
Privacy Note
Guardian classifies messages in real time to reach a verdict. It only writes a record to its audit log when it actually flags or acts on a message (suspicious or scam) — that record includes the verdict, category, reason, action taken, and a short snippet of the offending message for the mod-review card. Ordinary clean messages are never stored by Guardian.
Guardian works alongside Nova's other community tools — ZADAO Stats for engagement, and the daily community snack for conversation. Just ask Nova "what can you do to protect my community?" to get started.
Last updated