I agree with Maik that the agent is an excellent option but would suggest it comes with a downside (see option #2) Either way you are going to have to read the file system and the Agent and Hz Discovery can help but in different ways.
My thoughts on a possible solution...
Option#1 I'd imagine you would want to start with a process classifier that looks for Chrome running and if it is then you would have to do a multi-process step of gathering the username of the folks on the host then fill those into the custom pattern to gather the contents of this folder and reference them back to the chrome process. Rinse and repeat for approved browsers, Edge, Firefox, Safari, Brave.. ect ect..
Option#2 If you can't be convinced that you might catch a running browser then using a pattern extension on the compute find the users and look through the file system and relate it back to the computer. I am against this option as browser extensions do not support the operation of the compute but are more a component/attribute of the running (chrome/browser) process.
Either way, I think you have options. It's gonna be a little complicated and will probably have to use eval scripts in your custom pattern and with the agent, you are going to have to do some unique parsing and querying .. but when you get it to work, we'd appreciate you sharing it with the Community!
