I am working on a CPU emulator for a circuit simulator. In the CPU emulator, you can create instances of different CPUs and run machine code directly on them. One of the tools I created is an assembler to convert assembly code into machine code. This tool supports linking multiple code files together into one project (kind of like modulescripts). Because of this, code files could get very complicated. I am aware of filtering requirements for user-generated data, but considering I already compress the data into my own hexadecimal format, do I still have to run the text through the filter? It can only be decompressed by the user who wrote the code. The CPU emulator is already client-based as the server cannot handle multiple CPU instances without hogging up too many resources. I just need datastore to allow the user to more easily create multi-file projects.
Not entirely sure I understand everything you described, but if a user cannot view or obtain in some magical way other users’ raw string data then you don’t need to filter at all.