Hacker News Reader

Comment by morpheuskafka | Hacker News Reader

This is what we need app sandboxing for. No reason third-party apps should be able to read the browser's cookie database.

View on HN

As an aside, the Chrome cookies database on Windows is protected using the Windows Data Protection API[1], which ties encryption keys to a specific user. In the case of the Chrome cookie database, each cookie's payload/value is encrypted using a cryptographic key generated by the DPAPI which is only accessible to that Windows user. Of course, (and as is the case with most situations like this), this does absolutely nothing to protect users against malicious or intrusive programs running with the permissions of that user.

So yeah, you're right. App sandboxing please.

[1] https://en.wikipedia.org/wiki/Data_Protection_API

I thought the rule was: If you can see it plaintext on the screen, it's not safe.

Could you elaborate on that? Of course no computer is completely secure against all forms of attack, but I’ve found statements like these to not contribute very much towards solving any practical security problems.

On windows any program can read the contents of any window on the current desktop (with some exceptions, like UAC prompts and other windows that dim your entire screen). This has legimate use cases for screen readers or dictation, but of course it can also be abused. Same for X11. On more modern operating systems like Android the user needs to take very explicit action to allow an app to do that.

Wow, I had no idea this was also the case on Windows! I was aware of this on X11 and switched to a Wayland compositor as a result. Is this true for macOS? I know for sure that on more recent versions of macOS, applications which want to capture the entire screen require special user-approved permissions, but can they selectively read other windows?

The moment that a string of text appears on your screen, it is passed through between many different functions (and even pieces of hardware, with textures being stored on the GPU VRAM) and has unencrypted copies laying all around in memory as the application is running.

Any part of the space in which these copies exist being compromised will mean that your encryption is useless in the end.

You have to draw the line somewhere. Otherwise we'd be saying that HTTPS/TLS is not secure since the webpage is rendered to your monitor unencrypted.

I agree with the parent quote:

> I’ve found statements like these to not contribute very much towards solving any practical security problems.

But HTTP/TLS indeed is not secure from the point of your local machine. Its entire point is protection of data in transit.

One of the great things about Flatpak [1] is being able to restrict permissions for each application. Flatseal [2] is a useful GUI for controlling it. Zoom is available as a Flatpak [3]

1. https://www.flatpak.org/

2. https://flathub.org/apps/details/com.github.tchx84.Flatseal

3. https://flathub.org/apps/details/us.zoom.Zoom

Reply on news.ycombinator.com