Improving Clipboard security #84
Replies: 5 comments 5 replies
-
This is kinda related to X11Libre/xserver#197 (the ui prompt part mechanism), but its not part of xnamespace |
Beta Was this translation helpful? Give feedback.
-
This is such an amazing effort you put into this suggestion. This looks like a perfect blueprint of what to do going forward, but yes it is technically already an issue as @dec05eba mentioned. I would add this post to the issue that way it gets more visibility. |
Beta Was this translation helpful? Give feedback.
-
I personally don't like it:
|
Beta Was this translation helpful? Give feedback.
-
I think it can be streamlined a lot (from the user's perspective) with two reasonable assumptions:
So, if the window requesting to read the clipboard is the same as the one that set it, then read permission is always granted. Then I'd say only two confidence levels are actually needed: public (default), and confidential.
That would make for the default behavior, which should probably be configurable, maybe even per selection. A config file could look something like this:
For which window is which trust level, that too has to be configured, in some way. Maybe on top of XAuth, or as part of XNamespace. |
Beta Was this translation helpful? Give feedback.
-
Thank you for your contribution! We currently restructured the "Ideas" discussions and accordingly this discussion will be moved to the X11Libre 4 Good Ideas For Later category. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Improving Clipboard security.
This more like draft of draft and probably will do in some form in my free time on slowpoke speed, but it more like to start discussion if it is remotely sensible and I am not very familiar Xserver - thus i can totally miss logic.
I want start discussion from Clipboard security as it is look easier to implement proof of concept.
In 2025 is not malware problem main problem, but random proprietary software, which goes like "We value your privacy, our tool are enchanted by AI and in order to deliver 100% genuine experience we need send to our servers following information: all your keyboard keystrokes, all clipboard, periodic screen shots, network information (we will scan your network and maybe inform about vulnerabilities), your pictures from camera and so on."
The idea of security lies on 2 points:
The low effort example for waiting user approval to access clipboard
For start data_kind, can be:
Trust factor, probably variable with following values:
0 - Fully trusted - only data_kind=CONFIDENTIAL needs user approval
1 - High trust - CONFIDENTIAL and PRIVATE needs approval
2 - Normal trust - passwords too needs approval
3 - Lower trust - message to user if clipboard accessed when window is not focused. Obvious LOGS_ERRORS and PUBLIC is accessible
4 - Guest trust - only public clipboard data available (and probably need focus)
5 - Untrusted - any interaction with clipboard needs user approval
To implement
Any comments? It is sensible or just overcomplicated idea?
Beta Was this translation helpful? Give feedback.
All reactions