yanyannoodle34-debug/OpenBotMOBILE ? reverse-engineered prompt

Build me a local web app for AI coworkers that I can actually trust with real work.

I want each coworker to have its own browser, files, and tools, and I want to be able to watch what it is doing on its own screen. Every action should be checked before it happens and logged after, so I can review browser clicks, file changes, and anything it tried to do. Include a simple place to create and manage coworkers, start a chat with one, and see its activity and audit trail. It should run on my machine with Docker, use a database for data and logs, and let me connect my own model key instead of shipping one. Also make it flexible enough that I can bring in different kinds of agents, not just one framework.

If you need to look up current docs for anything, do that.