Request: The Most Important Meta-Feature

Perhaps the biggest reason someone in risk will want to use Haven is to get some hard evidence. I know, and have known for long, that someone is breaking into my house. I was again away for a week and now it has become in your face: we know that you know and let's see what you can do about it. Just before I returned (an hour or so before), someone entered and spilled water randomly in the kitchen (it hadn't even dried up). The only phone that I was able to configure with Signal had its charging cable taken out, so it stopped sending heartbeats after a couple of days. (This time, I didn't rely on power banks. I put each phone on the mains charging, with inverter support, so that power cut couldn't affect it. One phone had no logs for the whole week that I was away. One did, but doesn't seem to have registered anything, or else had it deleted. It seems to have gone zombie, so it is still registering logs, even though it is apparently disactivated. The third one had the charging cable taken out. I even have cameras installed, which were messed with and I am struggling to get them working again.). In such a situation, Haven is useful to me only if it can give me some hard evidence, otherwise I already have physical evidence, but I can't prove it to anyone.

Also, the new RC version that I installed, does not take photos no matter what, even if I wave in front of the phone and shake it up vigorously.

Please consider this as a meta-feature request, both with regard to the phone cameras and the sensors. Even the least bit of hard evidence that Haven can produce will be much more useful than lots of logs which can't be considered hard evidence.

Start by reproducing the reported RC-version camera failure and review Haven's phone-camera and sensor behavior. The issue names no files, tests, or entry points, so the scope needs clarification before implementation. Done would require reliable capture of verifiable evidence from the cameras or sensors, rather than logs alone.

Written by the indexing model from the issue text.