Very true, I managed to blow up a very expensive equipment during testing, more embarrassing I actually designed the equipment. I would run it from a power supply and take measurements with an o-scope and sparks would be flying in my lab. Very embarrasing. Would be nice if you could come up with a solution to the problem. 

The solution is obvious. 

1. The only universal solution I've seen and intend to use is the isolation transformer. Actually I built one and this sensor box is just the extension of it:

https://hackaday.io/project/9626-isolation-transformer

2. An other possibility to use differential oscilloscope probe, but this is costly and probably limit the bandwidth of you scope.

3. You can use mains isolated scope, like the battery powered, portable ones

4. I've an "unorthodox" solution to this problem in my mind, what is able to solve the problem in this special case, when you just want to measure the mains voltage and current of a device, but not suitable for internal measurements (like I did with that bloody blowing PSU). The concept is based on sensing the line and the ground of the incoming mains and cross-switching them if it is necessary. It can be achieved with some basic circuit. The cost far less than an isolation transformer, but also the capability and the safety. I just didn't build my idea because I already have an isolation transformer.

So, back to the sensor box. Please don't expect any voodoo on it. It is just a resistor divider a few shunts, switches and connectors. The solution is not here. Hopefully, it will be finished and documented by end of the week

Nice project

Big bumm happened to me before. Isolation is critical.

When I started to work on my (still unfinished) CNC drive, my power supply blown. When I tried to repair it, connected the oscilloscope to the primer side's rectified ground. The mains circuit breaker of the house went of, and the copper foil vaporized from the PCB, rendered the power supply to irreparable.