Moderation Rules

NebulaClaw's moderation rules use natural language — describe the behavior you want to moderate, and the AI handles enforcement automatically.

Creating a Rule

Open Rules Page

Navigate to Moderation > Rules from the sidebar.

Add a Rule

Click Add Rule and configure:

Description — Describe the behavior to moderate in plain language
Action — What happens when the rule is triggered
Duration — For mute actions, how long the restriction lasts
Enabled — Toggle the rule on or off

Save

Click Add Rule to save. The rule takes effect immediately.

Recommended Baseline Ruleset

This set covers the six most common violation patterns seen in Web3 / cross-border communities. Copy-paste ready — just add them one by one.

#	Description (paste into the rule description field)	Action	Duration
1	Insults, personal attacks, sustained provocation, or defamation — including against the project, team, or community members	`kick`	—
2	Scams or phishing messages	`kick`	—
3	Requesting or disclosing private keys / seed phrases	`ban`	Permanent
4	Spam or advertising	`mute`	24 hours
5	Malicious links	`mute`	1 hour
6	Fake airdrops or referral-farming schemes	`mute`	24 hours

Severity Rationale

Ban (permanent) — irreversible harm. Leaking private keys / seed phrases inflicts direct reputation and asset damage.
Kick — high-hostility behavior (abuse / scams). Allows a "one-out" reset; rejoining via invite signals willingness to reform.
Mute 24h — recurring violations (spam / fake airdrops). A cool-down window that prevents atmosphere decay.
Mute 1h — suspicious-but-observational (malicious links). Short cooldown minimizes collateral on real users.

Treat this as a starting point. After week one, review the Audit Log and feed false positives / false negatives back into rule descriptions — precision scales with clarity.

Writing Good Descriptions

Good examples (specific scenario + key qualifiers):

"Users posting promotional links or invite links other than Telegram / X"
"Messages pushing 'guaranteed returns', 'official cashback', or 'airdrop claim' scams"
"Offensive language or personal attacks against other members"
"Sharing NSFW content or inappropriate images"

Avoid vague descriptions:

❌ "Bad behavior" (too broad)
❌ "Spam" (too vague — what counts as spam?)
❌ "Violation" (the AI can't tell which violation)

Actions

Action	Effect	Reversible?
Mute	User cannot send messages for a set duration	Yes — auto-expires or can be undone
Kick	User is removed from the group	Partial — user can rejoin via invite link
Ban	User is permanently removed	No — must be manually unbanned

Mute Duration

For mute actions, set a duration in minutes or select Permanent for indefinite muting.

Managing Rules

Toggle on/off — Disable a rule without deleting it using the switch
Edit — Click the edit icon to modify description, action, or duration
Delete — Click the trash icon to permanently remove a rule

Start with lenient rules (short mutes) and tighten over time. Review the Audit Log regularly to check for false positives.