If you read Rossi's patent (Fluid heater) you realize there are some aspects that are a little more complicated than the simplistic way you see things. For example, you can read about a voltage modulation at claim 19:
[...] wherein said controller is configured to monitor a temperature from said temperature sensor, and, based at least in part on said temperature, to reinvigorate a reaction in said fuel mixture, wherein reinvigorating said reaction comprises varying a voltage of said voltage source.
A third party might not be able to get the desired effect from the E-Cat if it does not know how it should work. Moreover, if it were such a banal object, they would have replicated it everyone with the same results as Rossi. As for point 4, if you had a revolutionary discovery, would you really trust in an NDA? I believe that, like Rossi, you would like to show your invention personally or at least to be present during the test.
SSC - it takes a while for you to read what i say. I have said nothing about how simple of complex the controller in the black box should be - merely that it can be inside the black box making your point here wrong. If it has to be a human being continuously monitoring things then Rossi has no commercial nor even practically usable product. In most of the cases we know about, anyway, the devices worked fine and stably without human intervention for long periods. Human intervention seemed necessary during the GPT, but there is an alternate possible reason for that, as I'm sure you can understand.