Naming conventions for private sensors?

(For background see below)

I have developed a whole bunch of sensors for a device from companyx that I’ve written to be shared with other HA users. I’m trying to make the code/ sensors as clean as possible. There are a few ‘sensors’ I have had to create that users don’t need to see or use. If this was python or C I would have declared them private in a class or just used a naming convention like __ (two underscores) or $$ to make it clear they are internal. But I can’t do that with HA and YAML.

So what is the best/ preferred way to name sensors that:
a) users don’t need to see and should not change?
b) are derived data, ie something I have computed for the users ease of use, rather than a value from the device? (And I want to make it clear these are not values directly from the device)

Say the sensors are called:

sensor.companyx_switch1
sensor.companyx_switch2
sensor.companyx_switch1_temperature

I’m thinking private sensors could be (say)
sensor.companyxprivate_securitytoken
ie using companyxprivate as the prefix which will sort below .companyx_ when sensors are displayed.

And for sensors where the value is computed:

sensor.companyxd_combinedstatus
binary_sensor.companyxd_switch1_overtemperature

Using ‘d’ for derived

Any thoughts, please?

Background (how did I get here)
I’m working with the company that manages one of the devices in my home. They have an API that is aimed at corporate customers, however they have invited me to work with them on consumer access to their API. (So I/we can use it in HA :slight_smile: )

Ideally this would become a HA integration but as a first step to allow others to quickly make use of the API I have developed a whole bunch of sensors in YAML. (So I can just share the YAML, people can use it and we can work out what any integration should look like.) Also it would allow me and the company to get feedback on the API quickly rather than wait for the development of a HA integration. (Although having being working in Python this morning to surface something completely different in HA I think I might have saved myself a heap of time doing it all in Python from the start!)

Probably would be easier to do this as an integration and then either not expose the bits out as sensors at all. Or declare them as diagnostic (which you can’t do with manually created sensors only with sensors published from an integration). Diagnostic sensors do not by default show up in the auto-generated dashboards and have to be explicitly added to show up.