Basically, repeat the experiment under a wide range of conditions, and show that the conditions for success, if any, are far beyond the original claim. I always loved the ‘mythbusters’ approach: if one bible can’t stop a bullet, how about two bibles? ten? where is the cutoff between true and false?
That doesn’t answer the question, they asked if it’s open source. I agree, I don’t want to replace one sketchy data-harvesting service with another, I’d be a lot more comfortable giving shutup10 control of my system if it was on github or gitlab.