is a basic of all the scripts they were using and I came here today to ask if it’s possible to track if a player uses the keyword “;kill” and the script finds if that word is used in the players chat and it kicks them
Checking for keywords that let exploiters do their exploits is very band-aid. The exploiter is likely abusing a vulnerability in your remotes, so you will want to check for any remotes you are using. it could also be an admin commands system you are using that has vulnerabilities as well.
Exploits of this severity don’t just happen. There is a vulnerability somewhere, whether in your code or in code someone else inserted.
Maybe you don’t have a damage related remote… but do you have a remote that might destroy an object? Or maybe a remote that sets a property? The issue might not always be obvious, but instead an incorrect use of an otherwise safe remote.
There is no admin commands that contain ;kill the picture I provided is just an example of what commands they’re using, and also the picture I showed is when one of those words are chatted it kicks them, but the only way I can do it is when the play just chats ;kill I want it to kick the player if the chat has detected /kill anywhere in the chat message
Exploiters can take advantage of the remotes you mentioned that will destroy objects.
They can use the remote to destroy another player’s character head, which would kill them instantly, or disable them by destroying their humanoid inside the character.
You should put some security checks on the server, such as only destroying certain objects (I suggest using CollectionService).
Exploiters have many more functions than the command line or a regular local script. You might want to check how you are sending data through and be positive that it is not through any remote events.
As @Autterfly said, the vulnerability is somewhere. Exploiters have very tricky ways of bypassing various game anti-cheat/detecting systems.
Are you using free models? Your game may be backdoored and that’s how exploiters are killing people in your game. You should look at all of your free models and make sure there are no fishy scripts that are giving people access to the server-side. This is the culprit most of the time in games with severe exploit problems. It only takes a suspicious patch of grass to backdoor your whole game.
I would also recommend making sure all of your remote events are secured because it only takes one event that isn’t secured for exploiters to take advantage of. For example, for your guns, you should process the damage dealt and do all of the checks on the server-side, cast raycasts to make sure they could actually shoot them, etc. (You probably already have this done just reminding).
Also, detecting keywords in the chat is useless. 99% of admin systems for exploits have their own GUI that exploiters can use to trigger the commands without chatting anything. As I see you are using it to block people using bad language, you could instead use the Process Message module script in the chat modules to just filter the message so it is not sent. To top it off, many innocent players may get caught in the crossfire and say an admin command when watching a hacking in the game use one. (Might think they all got admin commands. I see people do it all the time).
Since exploiters are using a FE kill/fling, this can easily be prevented by checking the rot Y velocity on the server. If it above 200 they are probably using FE kill/fling.