Issues with Word Blacklist
Alright, so Earlier this morning I had some issues in my server. I was having to ban at least five suspicious and shady accounts after which some hateful people came on my server. Now during the conversation Waya detected a blacklisted word where there was no blacklisted words used on multiple occasions.
I have showcased fragments of my logs with the particular incident taking place in the snapshots bellow. I hope you can provide an explanation + a solution to this bizarre behavior of the 
62 Replies
yeah it's because of this. Waya filters out non alplhabetical words like "what! it's"
I am not 100% sure how I should be able to dedect such things because my intention was that @Waya filters the word(s) that have weird chars in it, for example
a**ss)hole it would read asshole
Okey I made a fix for it, I will push it once I have more updatesYes I figured already that that's what caused the issue. The update to this should be pudhed now knowing that I'm likely not the only one struggling with this issue.
Speaking of issues, I enabled anti link filter, but I'm unable to add links that I don't want on my server such as links to nsfw or dark web along with some certain server invites.
Also sorry for the delay with the reply, I fell asleep, but I'm up now.
anti link disables generally all links, besides of discord related links like âdiscord.com/termsâ (excluding invites) as this was a wish from someone a while ago. The fix will be deployed in around 1 hour
The fix has been deployed around 2 hours ago*
Awesome, thank you!
As for Anti Link feature: with this you should have taken the concept of the idea from such bots as mee6 and pro bot. The Anti Link feature on these bots allow users to customize what links are and are not allowed trough the blacklist and whitelist function. The lists basically are like modes where if you enable blacklist, then only blacklisted links will be removed and punished for where as with whitelist you'll be able to set specific websites and links that are permitted on the server, but everything else link related will be nuked.
yeah I might look into that. the thing is just that mod never realyl got my main attention as not really anyone uses it, reasons I never advertised it publicly anywhere
Hey @Luna, I was setting up the anti-scam feature, but I can't edit the time out length.
oops
huh
works fine for me?
Umm @Luna ? I think your update on the blacklist feature kinda...pretty much flopped. Waya is still blocking harmless words such as "contradiction" because it involves blacklisted word "dic"
Here is a fragment of our logs to showcase it. And yes, this is on the same server Kiki works on. We both run the said server along with one other admin, so I decided to drop by and leave this here for you guys.
As for the time out setup failure, I ran in to that issue over on my server whilst making my moderation system more strict and secure in case of raids or attacks. Given that one of my partner servers got raided and then locked down for maintenance.
the fix only included things like "what! it's". The problem is just the more things I add to prevent false-positives, it just gets easier for users to work around that blacklist
Well then add AI so the not can determine context. Shouldnt be hard to do
eh
you think making ai's is easy?
You can use existant AI and make the not work woth that for context deduction
you are aware that public ai's (like open Ai) cost money, and depending on the usage not that less. And from where should I get that money from? Waya is making negative monthly
Yes I would like to do that, but I do not have the resources for that
and no one says Ai's cannot have false positives
Auto moderation is just an addition to your current moderation and will never replace real humans
Well we can't have a not blockimg out harmless words just because its stupid and cant tell what's swearing and whsts not
Well Real humans dont sit around all day watching each channel forcall sorts of swears and racial slurs.
That's where the bots fill in
I am aware of this, but bots will never be able to replace humans, because (obviously) human brains are way more complex and can understand humor and context way better than code
well code can be programmed to understand humor and context close to how humans do
Bots are just here to help us, not replace us. And it's impossible to program any auto moderation system that works flawlessly. I have seen attempts with Ai's in the last couple of years and they all discontinued
Because either it's too expensive, not worth it or just not possible with reasonable affords. You have to understand that I am working alone here all day, I have no team or any company behind me
I already try my best
Well surely mee6, Carl bot, pro bot, sapphire, Dyno all do a better and more effective job at blocking out only actual swears and not harmless words and even come with preset swear libraries in more then one language in some cases.
because many if not all even bother to check if you put any characters between words like
ass*@holeI know you work alone and are trying your best, but seeing your struggles it's best you take Waya and your friend takes DisPing and you merge them in one bot and then form a team
This will not happen. As we both code in different styles and are not interested in the same functions. And I personally do not think he would be a great addition to this project
Well seeing his bot i think he would be a great addition to it
Working in a team often just slows me down
I use both of your bots across three servers each
And how would you know that if you always work alone?
Well I know his code and I know my code. We both are at different skill levels and working with ppl who do not follow the same standards as the other can harm the project and just slow each other down because everyone has to check what the other did before continuing being productive.
In addition I know him for not testing things in staging, using TypeScript (the programming language) in the intended way or generally using typings correctly if at all (which is the purpose of even using TypeScript)
Because I worked with others in the past on other smaller projects already, for example @Pronouns or @Auditional Text
And I do doubt he is even interested in doing that
Well Chill's bot works either way so he must be doing it right then. Your bot is also working but has issues that can be inconvenient to literally everybody.
there is a reason why moderation is not even being advertised
3%/3k have ever used mod (including ppl who dont use it anymore or never set it up correctly)
my main focus jsut isn't mod anymore since discord's release
I rather focus on things that are used by many
Cause no one wants to have harmless words blocked especially if the punishments for swearing is a mute/timeout.
So you have some options to fix the issue:
1. You find an affordable AI Or make one yourself and use that to upgrade moderation system to mostly rule out any false triggers and help it see the context.
2. You merge Waya with DisPing and allow KingChill to take a crack at solving this under your supervision.
3. You just fix all the filters by making the bot ignore swears hidden in other words and make it react if the swear is said directly and matches the blocked word.
how can ppl know that it doesn't work without ever using it, and even that happens rarely only in such situations...
here, everyone can fix it now
Well i know six servers using Waya for mod alongside other features and struggling with the blocking of harmless words.
I am in 18 servers with waya and only 1 server uses it. And I knew back then that for example
bypass is an issue.. but it's hard to fix such things without completely making it easy for ppl to bypass the word blacklistSpeaking of bypassing, you might wanna make sure Waya detects swears coming from bots and webhooks as I've had incidents where both were used maliciously and produced swearing and other junk
I mean yes, I have on possible solution:
instead of searching for (ie)ass(which dedectsbypass)
making it searching (ie)that would break all logging from every botass(with spaces), but makes it bypassable with (ie)youass
Small price to pay for safe and clean server.
That's likely the method most bots use and for the most part it does the job, so why not go for it.
then I will deploy this in 1-2 hrs
but also how can webhooks and bots send insults?
Excellent
People doing it trough them
if a bot has a say command, why do users have access to that. And if a bot is hacked, insults is the smallest problem. And if a bot is malicious, why is it on your server either way
Just hours ago u had one user used our translator bot to send words that are on way's Blocklist. Translator is a verified discord bot and is not malicious nor was it hacked.
A user simply used the translate feature or something like that
okey,.. can you give me the id of that translator bot (for testing)
Apologies for the delay, i was having dinner. Now here's the ID: 360081866461806595
yeah it doesn't look like it's even possible
yup, I cannot get the content of /command messages
I pushed it;
in this example,
hello is blacklisted:
(green: detected; red: ignored)Good morning Luna and thank you for pushing the update. I do apologize for coming off as rude, arrogant, childish, annoying or otherwise as a nuisance. I was pretty tired and i have an on going drama in my server that was brought in from another server.
Anyway, as you know, Discord has a interesting way of displaying things for mobile users (phones & tablets) and PC/Desktop users. I saw within my notification bar on my phone that you have used text coding to apply color to your text.
However as you can see in this snapshot showcasing your message from the perspective of a mobile user with a phone, the colors are not showing. This is an on going issue with discord that they're not fixing.
Surely there must be away to insure mobile and tablet users get to see and use text colors. But either way, for me everything looks colorless.
+: detects
-: ignores
wa reload --yarn:tsc
âą Reloaded all interactions in 21 seconds
added
Huh?
viewing the actual words in the list
lol
Oh yes that's a good idea
it's already deployed
All bots have it so itvwas high time you also had added that installment. That's good it's deployed.
lol
Morning đđ! I added Pronouns and Auditional Text bots to my server. Hope they come in to use for me.
As for Waya, i hope the new update improves the blacklist feature although it does leave it exposed to a flaw.
Good morning @Luna, I would like to know how can I remove a word or two from the black list. I have accidentally added something there that i did not want to add.
So now i need a way to remove that accidentally added query from the black list.
the same way you add words will remove them
removing doesnât support bulk tho as far as i remember
Do Elaborate
just click edit word list
and enter the word you want to remove
note you can only remove one at the time if I remember right, will add something later that it also supports multiple to delete