Being a UART protocol, it seems like physical access is assumed.
It's interesting as an asynchronous variant of g-code. Given the inherent safety implications, I think an improved protocol would benefit from some sort of checksum and feedback mechanism. Also, it seems perhaps a bit confusing to overload both position and velocity control using the same "l" code.
The security should be baked into the relay aka. whatever you use to translate the data from a IP based protocol to serial data.
T-Code is designed to be simple to understand, implement and decode by micro controllers over a wired connection and those micro controllers often lack the processing power to do cryptographic security. This relies more on physical security.
It's interesting as an asynchronous variant of g-code. Given the inherent safety implications, I think an improved protocol would benefit from some sort of checksum and feedback mechanism. Also, it seems perhaps a bit confusing to overload both position and velocity control using the same "l" code.