Comment by handoflixue
2 days ago
I think it matters a lot here if you can mathematically prove that it is "safe":
The only reason I'm comfortable transmitting sensitive over the internet is that we've got a bunch of proven systems like "encryption" that can guarantee my ISP, etc. doesn't learn much about my actual traffic.
Fundamentally, if you're giving an application your passwords, it is precisely so that the application can transmit those passwords, in full, to the relevant services. If you don't trust "mathematically proven safe" then it's already unsafe - and it's unsafe to a ton of third parties used to route the internet.
It's important because that means this is a harmless issue - there's a linked 3rd party mathematic proof by a credible authority on security (CloudFlare), and other password managers have similar features.
"I'd like clearer documentation" is still a reasonable request, but I don't think the OP is at all justified in acting like this is some outrageous trespass.
If you care about a harmless issue this intensely, you should be taking some responsibility to actually inspect the code. Given there was a feature flag to disable it, even a cursory scan with a cheap LLM probably would have found the "issue"
> Fundamentally, if you're giving an application your passwords, it is precisely so that the application can transmit those passwords, in full, to the relevant services.
This is about secrets in general. Those could just as well be used to encrypt local data or sign session cookies for the internal web page. And sending the secret to the service it is for is different than sending it (or at least a part of its hash) to another service the user might not trust.
Sorry, can you provide a link to where the HA application is scanning the filesystem for secrets in general? I thought this was just the passwords you hand over so it can automate things
I agree that it only applies to HA secrets but HA also has a web interface. Secret doesn’t mean only password
No matter how you encrypt the data, transmitting it unavoidably leaks the fact that you're running software that transmits that data. There's no way to mathematically prove this is harmless, because there's no way to know how that information could be used against in the future in ways you didn't predict.
No software should ever make network requests without your consent.
Yeah except the web browser and mail client I wish the default would be no open sockets for you ever.
Unless it is a distribution policy to respect privacy by default it is just to much work to keep up with all the leaks the system might have.