Quick Note!
The auto-mod filters are crafted by the developer and do not allow inputting custom regular expressions or inputs, at least at this time. However every filter is configurable. You can configure the following properties:
Punishments (nothing, timeout, kick or ban)
Patience (amount of verbal warnings before punishment, if any)
Time windows (for time-based filters: how many seconds between messages, etc.)
Thresholds (how many emojis, duplicate letters, etc. are considered spam)
Offensive Words
This filter matches some very common offensive/derogatory words and racial slurs. It's fairly robust in the sense that it will match letter and emoji replacements like 🇳, arbitrary spaces or repeated letters.
Word filters can always be defeated with little to no effort but this will catch the majority of common uses, if someone tries to defeat the filter in your chat they typically won't last long anyways.
Messages deleted when caught
Invite Links
This filter matches any Discord server invite links, like https://discord.gg/xmPMufUF9N. (That's the Discord server for the bot!)
Messages deleted when caught
Malicious Links
This filter matches an internally maintained list of links that are encountered by the developer and moderators in the development server by random, usually malicious users. It's not exhaustive, but any new links encountered are added to the bot pretty fast.
It includes links like IP grabbers, redirects into fullscreen are you sure you want to leave this page max volume jump scare and phishing (scam) links that are slight misspellings of official websites like Steam or Discord.
Messages deleted when caught
Duplicate Spam
This filter matches messages that repeat letters or phrases like "omg omg omg omg omg omg omg omg omg omg", "aaaaaaaaaaaaaaaaaaaaaaaaa", or repeated empty lines.
Large messages deleted when caught
Message Spam
This filter matches messages that are sent in quick succession, like 10 messages with less than 2 seconds between each message.
Capital Letter Spam
This filter matches messages that are at least a reasonable size and consist of roughly >70% capital letters.
Scam Detection
Experimental: This filter attempts to detect scam bots/compromised accounts that send malicious links like “free nitro giveaway! <insert scam website here>” in your server. It's marked as experimental as it may have false positives and will likely need refinement.
Messages deleted when caught