Hi, this post is a braindump for me of a certain design decision in my Discord bot, Lucinne. For those who are unfamiliar, Lucinne is a Discord bot designed for Rosa's TF Zone to support some light, kinky fun by restricting permissions and replacing messages. I've had a few "design philosophies" around creating it:

• Safeword comes first.
• The different effects of the bot should be unique, but not overly-specific.
• Lucinne is a roleplay assistant "with teeth".

"Safeword comes first" means that everything should be cancelable/disabled if you react with the server's ":safeword:" emoji. I've put a lot of effort into making sure that the safeword works for all of the features. There are rare exceptions, but those are for specific individuals who ask for it and are given explicit permission by the server owner to opt-out of being able to safeword.

"Unique, but not overly-specific" is slightly harder to explain without using an example. The "squeak filter" is straight forward. There is a % chance of each word in the afflicted user's message being replaced with a sqk sound. There's also an optional requirement to include the word "Smile!~" in the message or else the entire thing will be replaced. So random chance and must-include-word as mechanics. Compare that to the "feral filter", where every message is totally replaced with dog/cat/cow noises. Others can vote to reveal your message or keep it hidden. There's no random chance anymore and it doesn't matter if you include certain words or not. There's minimal overlap between what they do and it makes each filter seem more unique and fun as a result.

"A roleplay assistant with teeth" comes from my experiences with other bots or tools, and from my own preferences. There are some tools that will do little more than change your name. I think Muzzed Fox and Enforcer are the only two Discord bots that do more - mostly by offering a gag-speak option to scramble your words. When I say Lucinne "has teeth", I mean that it has bigger features that do more. Object filter will reduce your ability to interact down to saying ".....". Robot filter will limit you to only pre-made messages. Feral filter will replace any image you send with paws. I haven't seen any other bot offer the leashing mechanic that Lucinne does. I want everything Lucinne can do to have weight to it. I want it to be more interesting and interactive than what you could do with only text. I want you to feel the surprise as you randomly let out a sqk, and to feel the need to Smile!~ every time to send a message.

So back to the title of the post: why do I only allow self-inflicting the squeak filter options? It's a mix of keeping Lucinne's teeth and keeping the safeword as a safeword. It's hard to explain but I'll try my best.

I want to make one thing very clear: you should always feel comfortable safewording when you need to.

• If you can easily opt out of the bot for a single message, it reduces the weight of the bot's features.
  • People will use it if it's easy to do. I've seen it happen before when I used Lucinne in another server.
• Safewording with the ability to immediately re-apply that filter is nearly the same as a single-message-opt-out.
• Safewording to send a few messages of casual conversation then un-doing it is a very casual use of the safeword.
• Casual use of the safeword lowers the weight of the safeword.
• Lowering the weight of the safeword is dangerous because it should always be a clear indicator of "I do not want this and it must stop". I want to make one thing very clear: you should always feel comfortable safewording when you need to.

Squeak filter was the first filter-feature that I made for Lucinne. It has self-inflictability mostly for historical reasons and because it is the most casual in the set.

Let me finalize this by saying that if Rosa asked for it, I would add the option to allow self-inflicting things and even single-message-opt-outs. The bot is primarily made for her server and I appreciate that she has let me host my bot there. But I personally am strongly against it.