Note: The following is just my interpretation of the rules.

In article , Leo Szumel
writes:

If you don't use a specified code, you must identify using a specified
code.
For example, if you design your own protocol (unspecified code), then you
should design the system to ID every 10 minutes, or every transmission.

That should not be a problem. I envisage we would use an unspecified
code for our data transmissions, but we could self-identify with an RTTY
broadcast every 10 min. This will all be computer-controlled so that
should be easy.

I think there's a problem with using a code that is not publicly available. ID
is not enough; if the message cannot be read by a suitably-equipped monitoring
station (read: FCC) what you have is a form of encryption.

Amateurs are not allowed to intentionally encrypt or otherwise conceal
transmission meaning or content, with one exception: remote control commands.
So the "turn off" command would be OK to encrypt, but not the data coming from
the remote sensors.

73 es GL de Jim, N2EY