Documentation Index
Fetch the complete documentation index at: https://mintlify.com/onesoft-sudo/sudobot/llms.txt
Use this file to discover all available pages before exploring further.
Moderation Rules are a declarative filtering system that lets you define conditions SudoBot checks against every incoming message or user profile, and specify what actions the bot takes when a condition is met. Rules are evaluated in order, and the first matching rule wins by default.
How rules work
Rules live in the rule_moderation object inside your guild’s entry in config/config.json. Each rule in the rules array is an object with at least a type and an actions field.
{
"YOUR_GUILD_ID": {
"rule_moderation": {
"enabled": true,
"rules": [
{
"type": "word_filter",
"words": ["badword"],
"actions": [{ "type": "delete_message" }]
}
]
}
}
}
Rules are order-sensitive. The first rule whose condition matches is executed. Subsequent rules are skipped unless the triggered rule has "bail": false.
Common rule attributes
Every rule, regardless of type, supports the following attributes:
| Attribute | Type | Default | Description |
|---|
type | string | — | The rule type (required). See Rule types below. |
name | string | — | Optional identifier used in bypass references. |
enabled | boolean | true | Set to false to skip this rule without removing it. |
mode | "normal" | "inverse" | "normal" | In inverse mode the rule triggers when its condition is not met. |
bail | boolean | true | When true, stops processing subsequent rules after this one fires. |
actions | ModerationAction[] | — | Actions to take when the rule triggers (required). |
for | MessageRuleCondition | — | Extra condition that must be true for the rule to apply. |
exceptions | MessageRuleCondition | — | Condition under which the rule is skipped entirely. |
is_bypasser | boolean | false | When true, this rule bypasses the rules named in bypasses. |
bypasses | string[] | null | null | Rule names that this bypasser rule skips. Only used when is_bypasser is true. |
Rule types
Word filter
Triggers when the message contains a blocked word (whole-word match) or a blocked token (substring match).
{
"type": "word_filter",
"words": ["badword"],
"tokens": ["badtoken"],
"normalize": true,
"actions": [
{ "type": "warn", "reason": "Automatic: Blocked word" },
{ "type": "delete_message" }
]
}
| Option | Type | Default | Description |
|---|
words | string[] | — | Whole-word matches. The word must be surrounded by spaces or message boundaries. |
tokens | string[] | — | Substring matches. The token can appear anywhere inside a word. |
normalize | boolean | true | Normalizes the message and patterns before comparison (lowercases, removes diacritics). |
Regex filter
Triggers when the message matches any of the provided regular expressions.
{
"type": "regex_filter",
"patterns": ["disc(o|0)rd\\.gg"],
"actions": [
{ "type": "warn", "reason": "Automatic: Discord invite link" },
{ "type": "delete_message" }
]
}
patterns array accepts either plain strings (treated as regex patterns with no flags) or [pattern, flags] tuples, where flags is a JavaScript regex flag string such as "i" for case-insensitive matching.
Anti-invite
Triggers when the message contains a Discord server invite link.
{
"type": "anti_invite",
"allow_internal_invites": true,
"allowed_invite_codes": ["AbCdEf"],
"actions": [{ "type": "delete_message" }]
}
| Option | Type | Default | Description |
|---|
allowed_invite_codes | string[] | [] | Invite codes that are exempt from the filter. |
allow_internal_invites | boolean | true | Permits invites that point to the current server. |
Domain filter
Triggers when the message contains a URL whose domain matches the blocklist.
{
"type": "domain_filter",
"domains": ["example.com"],
"scan_links_only": false,
"actions": [{ "type": "delete_message" }]
}
| Option | Type | Default | Description |
|---|
domains | string[] | — | Domains to block. |
scan_links_only | boolean | false | When true, only text that looks like a URL is checked. |
MIME type filter
Triggers when a message attachment has a MIME type in the blocklist.
{
"type": "mime_type_filter",
"data": ["application/x-msdownload", "application/octet-stream"],
"actions": [{ "type": "delete_message" }]
}
data array contains MIME type strings such as "image/png" or "video/mp4".
File extension filter
Triggers when an attachment’s file extension is in the blocklist.
{
"type": "file_extension_filter",
"data": ["exe", "bat", "ps1"],
"actions": [{ "type": "delete_message" }]
}
Repeated text filter
Triggers when the message contains an excessive number of repeated characters or words.
{
"type": "repeated_text_filter",
"max_repeated_chars": 20,
"max_repeated_words": 15,
"actions": [{ "type": "delete_message" }]
}
| Option | Type | Default | Description |
|---|
max_repeated_chars | number | 20 | Maximum number of identical consecutive characters allowed. |
max_repeated_words | number | 15 | Maximum number of identical consecutive words allowed. |
Mass mention filter
Triggers when a message contains more mentions than the configured limit.
{
"type": "mass_mention_filter",
"max_mentions": 10,
"max_user_mentions": 8,
"max_role_mentions": -1,
"actions": [
{ "type": "warn", "reason": "Automatic: Mass mention" },
{ "type": "delete_message" }
]
}
| Option | Type | Default | Description |
|---|
max_mentions | number | 15 | Total mention limit (users + roles). |
max_user_mentions | number | -1 | User-only mention limit. -1 disables this check. |
max_role_mentions | number | -1 | Role-only mention limit. -1 disables this check. |
Image filter
Scans text inside images using OCR and triggers if blocked words or tokens are found.
{
"type": "image_filter",
"words": ["blocked"],
"tokens": ["tok"],
"inherit_from_word_filter": false,
"scan_embeds": false,
"actions": [{ "type": "delete_message" }]
}
inherit_from_word_filter to true to reuse the word/token lists from the first word_filter rule instead of defining them again.
Embed filter
Triggers when an embed’s text content contains blocked words or tokens.
{
"type": "embed_filter",
"words": ["spam"],
"tokens": ["buy now"],
"actions": [{ "type": "delete_message" }]
}
Profile filter
Scans the user’s display name, username, and bio when they join or update their profile.
{
"type": "profile_filter",
"words": ["blocked"],
"tokens": ["tok"],
"regex_patterns": ["disc(o|0)rd\\.gg"],
"normalize": true,
"actions": [
{ "type": "warn", "reason": "Automatic: Profile contains blocked content" },
{ "type": "kick" }
]
}
AI scan
Uses the Google Perspective API to score message content across multiple harm categories.
{
"type": "ai_scan",
"toxicity_threshold": 0.75,
"identity_attack_threshold": 0.5,
"insult_threshold": 0.8,
"profanity_threshold": 0.5,
"threat_threshold": 0.5,
"sexually_explicit_threshold": 0.5,
"flirtation_threshold": 0.5,
"actions": [
{ "type": "warn", "reason": "Automatic: AI-detected harmful content" },
{ "type": "delete_message" }
]
}
0.5. Set a threshold to a higher value to allow more latitude, or lower to be more strict.
File filter
Blocks specific files by their SHA-512 hash, regardless of the filename.
{
"type": "file_filter",
"hashes": {
"d3d0bfbe67707d003ab937212ee96309b7f7beb6871391064917b70c20fa5a67": "image/png"
},
"check_mime_types": false,
"actions": [{ "type": "delete_message" }]
}
hashes object maps a SHA-512 hex string to a MIME type string (or null to skip MIME validation). Set check_mime_types to true to also verify that the attached file’s MIME type matches the value in the hash map.
Example: combined ruleset
{
"YOUR_GUILD_ID": {
"rule_moderation": {
"enabled": true,
"rules": [
{
"name": "staff-bypass",
"type": "word_filter",
"is_bypasser": true,
"bypasses": ["invite-blocker"],
"for": { "roles": ["STAFF_ROLE_ID"] },
"actions": []
},
{
"name": "invite-blocker",
"type": "anti_invite",
"allow_internal_invites": true,
"actions": [
{ "type": "warn", "reason": "Automatic: No invite links" },
{ "type": "delete_message" }
]
},
{
"type": "mass_mention_filter",
"max_mentions": 8,
"actions": [
{ "type": "mute", "duration": "5m" },
{ "type": "delete_message" }
]
}
]
}
}
}
