All that stuff looks fun, but I'm utterly terrified at the idea of it malfunctioning. Like, in a false-positive way. And, as a professional deformation, I guess, it is basically an axiom for me that any automation will malfunction at some point for some ridiculously stupid and obvious-in-the-hindsight reason I absolutely cannot predict right now.
I mean, seriously, it isn't a laughable idea that a bomb that will explode unless you poke some button every 24 h might eventually explode even though you weren't incapacitated and dutifully pressed that button. I'm not even considering the case that you might have been temporarily incapacitated. People wouldn't call you paranoid if you say that carrying such bomb is a stupid idea.
I totally see where you're coming from, and I agree—this project definitely isn't fool-proof. But honestly, it feels like the best option for making sure you're really gone while keeping privacy in mind.
As technology advances, we will develop more effective means of determining whether someone is truly deceased, for e.g. something like Neuralink could provide significantly improved methods for verifying actual death.
It's completely open-source and you can self host it.