Earlier quoted context omitted.
I'm not the author but i can probably answer some of these: > * Most people do not have passwords in plain text — an SSH key protected with a passphrase is not "plain text", for instance I think you'd be surprised how many SWE do actually have plain text secrets. For example, if you do everything correctly with AWS auth, you still have a plain text token in ~/.aws -- it's short lived (typically 8hrs IIRC), but it's s…
Your local development machine will have the source code or images or IaaC configuration that you push to GH/AWS — a rogue agent can just modify that and wait for you to authorize the next push. I'd also consider short-lived tokens "in-memory during use", even if 8h is not very short. If you have AWS keys but do not have anything you use them for, then yes, a tool like this will help. However, the tool seems to imple…
Yes, but that's then concerning yourself with the remote state and thus the opposite of the sentiment you made when you said:
> I find that I worry more about my local data compared to my remote data
> I'd also consider short-lived tokens "in-memory during use", even if 8h is not very short.
I never said 8hrs was short. And once again, nobody is talking about in-memory use. If that's a legitimate risk you need to cover then VMs or literal air-gapped systems are your only solution. But for clear text in-memory to be a risk you'd then need to be running untrusted executables are root. Which then raises the legitimate question of "why the f*k are you doing that on your main development machine in the first place?!!"
> However, the tool seems to implement per-provider handling, meaning large surface area and new risks replacing old ones. Even standard, trusted crypto libraries are frequently misused in insecure ways.
It sounds more like you're commenting without researching this tool.
Scepticism is definitely a valuable asset when it comes to security. But what you're doing here isn't scepticism. It's just dismissive and presumptuous. The good points you're trying to make are undermined by the really obvious flaws in your assumptions.
> I can certainly explore the code, but as it is a "Show HN", I'd like the author to show it ;)
They have. It's literally the Show HN link to the source and it includes a fair amount of documentation to support the product. The questions of trust you keep raising can only be proven via an independent audit of the code, for example, yourself reading through it.
> My point is to understand clearly where does author believe this tool has merit.
That's right there in the README. And I think you already know this because you touched on another solution to the same problem when you discussed containerisation. Which then brings me back to my earlier point about constructive scepticism vs lazy dismissal.