I'm using KDE Plasma with Dolphin on Debian 12 Bookworm, and I want to completely disable all tracking and history features related to: - Recent Files - Recent Folders - Clipboard History - Search History - Recently Used Applications I want also to clear any data and history that has already been collected by these features. How to prevent KDE from re enabling or recreating those logs. Im looking for a thorough and systemwide solution that ensures KDE does not track or show anything from past activity. What is the correct way to achieve this?

I have Googled this one and tried so many combinations, with no success. I need the GDM login greeter to display one and only one username. In effect, what I need is something like this in **/etc/gdm/custom.conf**: [greeter] ExcludeAll=true Include=guest This does not work, however, because ExcludeAll is not a valid option. IncludeAll=false does nothing either. I've tried every combination that I can think of. Despite being discouraged, I modified **/etc/gdm/gdm.schemas**, trying a variety of combinations similar to the above. No go. I am aware of: gconftool-2 --config-source xml:readwrite:/etc/gconf/gconf.xml.defaults --direct --type bool --set /apps/gdm/simple-greeter/disable_user_list true but this is not an applicable solution, as it removes the user list completely. I need one specific username to appear. The only practical solution I can find is to, upon user login, run a script which will append that username to the "Exclude" list in **/etc/gdm/custom.conf**. This is hacky, and not preferable. This is for a public-use machine. People can log in with their LDAP accounts if they have them, or they can select "Guest" from the login screen. The people who log in with their LDAP accounts should NOT have their names visible on the login screen, but the "Guest" option must remain. Using gnome 2.28.

I am trying to configure a NetworkManager connection for IPv6 privacy. My goal is to assign a temporary IPv6 address with limited lifetime to my device. I manually crafted a xxx.nmconnection configuration file according to general NetworkManager documentation. However I bump on serious problems.

Address generation

The file contains several address definitions: ~~~ addr-gen-mode=stable-privacy # A address1=2xxx:yyyy::zzzz # B address2=fe80::1234 # C ip6-privacy=prefer-temp-addr # D ~~~ Request A is there to generate a random stable address. Request B creates a global manual address. C is a link-local address to reference "comfortably" the machine on the LAN. D requests a global temporary address. When I emit ip -6 addr show, the command only displays B and C. I could understand that A and B conflict (because I force a manual address) but documentation does not mention such a conflict. I checked with sysctl -a -r addr that the parameters have the correct value in the kernel. Despite many reboots, I could not get the stable-privacy address.

DHCP workaround

I thought I could replace the temporary addresses by a DHCP lease. This is a surrogate but I can't control the lease lifetime, probably because the lease is controlled by the box and the interface does not provide access to these parameters.

Fedora issue ?

I read here that Fedora "does not enable" temporary addresses. However I could not find hints about hard-coded or configuration-controlled disable. In the latter case, I should get temp-addresses after toggling the parameter. It is also possible that I missed a procedural step. I expect the temp address to be generated when I activate my connection. But it does not happen.

Help

Could you guide me in enable IPv6 temporary addresses in my Fedora 41 computer?

My host is Arch Linux and I'm using virt-manager to run VMs. I noticed it had access to my microphone when Cortana started listening to me after a Windows guest install. I've looked through all the settings in the virtual hardware details section but I can't find anything relating to microphones. There is a "Sound" device added automatically, and when I delete it the VM looses access to the microphone, but it also cannot output any sound. I'm looking for a way to cut off microphone access while still allowing audio to come out.

If one installs the Github Copilot software for the Vim editor, it *enables* the co-pilot by default, which could be seen as a privacy issue. I want to _disable_ the co-pilot by default and only enable it manually via :Copilot enable in the editor when needed. How may I do that?

I am connecting with web browser (Chromium, Firefox) using socks5 proxy via local port opened by ssh -D (currently openssh 9.8). Sometimes, I experience problems from the service in the other end, which do not occur otherwise. Is it possible to say something general about whether the fact that an ssh proxy is being used and be fingerprinted and detected?

I am considering including a copy of my /usr/share/dict/words file in a public GitHub repository for a project that requires dictionaries. Is this a bad idea, and if so, why? I'm particularly interested in the privacy/security (or even legal?) aspects. Are there common programs that add words to this dictionary, for example if I choose "Add to Dictionary" in a spell checker? Is the file likely to contain any sensitive information, such as my username (I checked that, and it doesn't, but there could be similar things I didn't think to check). It'd be impractical to look through all 104,334 words. Perhaps it's just the usr in the path making me unnecessarily concerned. I've read over these questions about where the words come from. However, is it probable that any words have since been added or removed? I suppose if nothing has changed, I could just get the source. But if some programs added helpful (non-personal) words, I'd want to keep those. In case it's important, I am running Ubuntu 23.10. But I'd prefer a slightly more general answer, if possible. ### Note I am fully aware that - it would be possible to point to the file path in code rather than "hard coding" it into the repo, and - this may not be the best free English word list . However, I'm not interested in using a different list *instead* of this one (in such a case, I'd rather just use both). And if I use a list, it's necessary that I can include the actual file.

Why are these things enabled by default on Debian12 with KDE (Plasma 5.27)? Maybe it could be useful for a few users if it was installed but not enabled and not running by default and just an option one could enable in the configs or which could be enabled by the user through a prompt. If it's running by default this also means that after upgrades it could be running again. This is a privacy issue, an undesired bloat service, and a larger attack surface even if there was a proper and vulnerability-free permissions-management for GPS-location-access. So **why is it enabled and running by default** (links such as existing KDE discussions about this and explanations please). **Is this the correct best way to disable it** (one may still have to check after every OS upgrade and this should also work for users who don't want to uninstall things if they use geoclue sometimes on demand): 1. sudo start-stop-daemon -T --name gpsd 2. sudo rm /etc/xdg/autostart/geoclue-demo-agent.desktop 3. sudo apt-get remove geoclue-2.0 (if not needed at all) Issue about it is here .

I'd like to set up an instance of Thunderbird on a Fedora VM (with a VPN). I want it to use ONLY an alias I have set up with the email providers. But I will still have to use my real credentials to log into the server. So they are in `.thunderbird` somewhere. I have to deal with websites on that VM that are pretty invasive by nature (rental tenancy applications like 2apply, snug.com, etc). Each are in a Firefox Container. They all have the facility to browse your system to upload files and documents, or 'drag-and-drop' - i.e. I have enabled those scripts in NoScript, and I have no choice about that. I'm worried that information will leak from Thunderbird to the websites, particularly my real email address. I understand/believe/have heard that linux doesn't allow programs to work without explicitly commanding them to do so, but... haven't I done exactly that with the websites by enabling file-uploading scripts? How should I think about this? Is it really a risk?

I just wanted to raise an issue (ie. report a bug) in some software package hosted on gitlab. I have a github account, but not a gitlab one. Whenever I log in from a new device (github), I need to "authorize" the new device. Now, I tried "logging in" on gitlab using the offered github account/password authentication. I got the usual "confirm this email address" thing. I did that. Then it wanted my phone number! At this point I gave up. Do they also need my social security number, date of birth and passport number? Maybe my mother's maiden name? What's the need for all this two-factor-authentication level security?

When using SSH key authentication, all available keys are tried against the server. Is there a possibility to set some SSH keys in a way, so that they are only disclosed to a given remote server list, for privacy? I've been looking but wasn't able to find any information or documentation regarding this at all. Thanks in advance

I have a set of SSH keys to login to accounts and servers. However, ever since I moved my main work profile to a USB drive (for sync and ease of use among other reasons) I have been unable to *directly* use the SSH keys in the profile for authentication, having to create local copies instead. Normal SSH workflow usually goes ssh-add $HOME/.ssh/id_server_key ssh username@domain Which works without problems. However, to prevent each machine from having a copy of my keys I moved them to an xexternal drive, USB formatted as vfat. Then workflow *should* become like this: ssh-add /run/media/myself/USB/.ssh/id_server_key ssh username@domain ...Which doesn't work. SSH complains about doing something to protect me:

@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@         WARNING: UNPROTECTED PRIVATE KEY FILE!          @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
Permissions 0644 for 'id_server_key' are too open.
It is required that your private key files are NOT accessible by others.
This private key will be ignored.
bad permissions: ignore key: id_server_key
Permission denied (publickey).

Which, while I can understand, it's not *really* protecting me in this case. The USB drive automounts in my distro, with a fmask of (I think) 022. I could unmount and remount the USB drive with a fmask of 077 but that would (1) require superuser privilege each time and (2) affect every single file in the device, not only the SSH keys. I've tried creating a symlink to the key, even with previous use of umask as I've seen in some tips&tricks: cd .ssh umask 077 ln -s /run/media/myself/USB/.ssh/id_server_key But the permissions can not be used this way for a symbolic link, so I get nothing. So, what **options** do I have here available if the goal is to **not** have a physical copy of the key on the machine? One of the machines is a family laptop that goes to various places, so I have an interest in avoiding some of the keys being discovered. So far I've considered the following: * Is there for example an option to force SSH (ssh-add primarily) to accept the keys from the external device? (I presume there should be - most programs follow a "I-know-what-I'm-doing" flag, but so far I can't find it in the manpage. * Setting a bindfs over the .ssh/ directory, but the problem is that the directory is non-empty (as it contains known_hosts and other data). * fuse-zip-ing the keys over /dev/shm seems to be a possible path as well. **EDIT**: As per @Gilles 's comment, yes, bindfs solves the issue. After a checking of the manpage, initial testing succeeded using the following invocation: bindfs -n -r -p 0700 -o nonempty /run/media/myself/USB/.ssh/ ~/.ssh Which binds disallowing access from other users (apparently even from root!), read-only, with reflected 0700 permissions. That reflection was the part that I lacked when working with bindfs originally. The nonempty flag was needed because of preexisting files such as known_hosts. This preserves the original workflow exactly except for an extra warning messae about not being able to add information to known_hosts (which in the end I might not mind). Of course, I'm bindfs-ing to another directory now (in /dev/shm) as originally suggested to make things slightly easier. I'm not as conspiranoid so as to encfs my USB drive (yet) so privacy-wise this is good enough for me.

For security and privacy reasons, I want to stop logging logins, as well as hiding information about currently logged-in users. So far, I tried revoking reading permissions to utmp and wtmp: sudo chmod o-r /var/run/utmp sudo chmod o-r /var/log/wtmp Which successfully stops w and who commands from working. They return without information about other users. However, as demonstrated in the answer in https://unix.stackexchange.com/questions/734857/other-users-are-logged-in-how-does-the-system-know-where-is-the-information-l it is still possible to access the information about other logged-in users via a DBus call to systemd-logind or elogind. How can I disable this specific DBus call? Or, in general, how can I prevent normal users from being able to look up other logged-in users?

I am providing photos related to a legal matter to a lawyer and she is going to give them to law enforcement without giving them my name (under attorney client privilege). I am not sure if she is tech savvy so my plan is to give her a physical USB stick with JPG files on it, organized into a folder structure. I want to make sure that my identity is not discoverable from the stick or the files. I use Ubuntu 22.04.1 LTS. I am going to use exiftool to remove the metadata from the images. What can I do to make sure that no traceable data specific to my PC is left on the USB stick?

I have been using Linux & Unix for a long time. The first time I learned Unix is in a computer classroom with many students, where the instructor told us that you can use ssh to log in to remote workstations provided by the university I was studying in. In that context, I immediately know that Unix is a multi-user operating system because there were more students than workstations. (Pigeonhole principle.) Somehow after many years of use, I know basically what the umask (default to 022 for write permission to the user, 002 for write permission to groups of users such as teams) is and that the default permission of a user (and a new user) is 755 for folders, 644 for files. But after thinking a little deeper about it, it seems that the default permission setting is insecure in such a multi-user operating system. Although it makes sharing data easier, I still feel strange when you can easily list & copy other users' files within their home directory. So is why the default umask 002 or 022 in many Unix systems? Is there any historical reason that *traditional* Unix chose this design? (One possible reason I can think of is that, during 1970 ~ 1980 computers are used mostly in scientific research and the computing power is very limited. So when the early version of Unix comes out, they decided that the default permission should be for easy sharing instead of privacy protection.)

**TL;DR** *If I am connected to one network and then connect to VPN it works fine, but if I then connect to a different network then the VPN disconnects and tries to reconnect over the new network. I need it to stay "connected" while switching so I can maintain privacy and have mostly-unrestricted browsing using public wi-fi.* Currently using Kubuntu 23.04 (laptop) and SteamOS (Steam Deck). I frequently use my school's wifi, since it's free and faster than what I can afford at home. The kicker is that a lot of websites and downloads are blocked (things like .io, gaming, Steam, social media sites, etc.), also they monitor everything you do. To get around this I figured out that all I need to do is connect to a VPN, however, the school's network prevents you from connecting to one. So I found a solution: first I connect to a VPN using a mobile hotspot then switch over to the school's wifi. Somehow the VPN remains on (this works with Android and Windows 10/11). But with Linux this does not work because it's like the VPN totally turns off/disconnects between networks and tries to reconnect using the school's wifi, which doesn't allow you to use a VPN...so it doesn't connect... I've tried using different VPNs (Nord, Proton, Open) and different setup options (Proton's app, Nord via CLI, Nord and Open via configuration files) and even different distros (Kubuntu, Lubuntu, and Mint) but haven't fixed the problem. Also, I do not have killswitch on. Is there anything I can do to replicate the way VPN connections work on Windows and Android? That way it remains on when switching between internet connections.

I am no fan of the "recent files / recent apps / recent places" features throughout modern operating systems. In particular, I do not want/need - **jump lists**, that show recent files opened with particular applications (e.g. Okular, VLC Player, LibreOffice, ...), - a "**recent places**" tab in file explorers showing which folders I have browsed in previous sessions, or that open up at the place where I was when I last closed an explorer instance, - a "**recent files**" tab in file open / file save dialogs, showing recently accessed / saved files of *any* program in *any other* program, - "**recent apps**" feature that track, which programs I use. First and foremost these features are troublesome for me from a privacy perspective, both in a private and a professional context. But they are also in the way of my personal workflow: I know which apps and programs to use for the task at hand and which files I am going to create, open and save at which places. Any proposition of some heuristic of what I *may* want to use comes as a confusing distraction. So I figured that - on a Linux OS, as compared to e.g. Windows - it should be easy to disable these features, in my case Manjaro (with KDE Plasma), but it turned out to be much harder than I anticipated. Here is what I did (I translated many of the items from my non-English setup to English, so they may not be the exact item names as in an English setup, but should be close enough): - In the Manjaro KDE system settings under 'Workspace/Desktop Behavior/Recent Files' I switched 'Remember opened documents' from 'All applications' to 'Do not remember'. This had no effect that I noticed. - Again in the system settings, under 'Workspace/Desktop Behavior/Activities' I opened the settings for 'Standard' (my only activity profile) and switched on the setting 'Privacy: Do not record the usage of this activity'. This disabled the files and places stored under 'recentlyused:/files/' and 'recentlyused:/folders/' and along with them the 'recent files' part in the jump lists when right-clicking opened programs on the taskbar. So why is there a (presumably "global") setting in 'Workspace/Desktop Behavior/Recent Files' if I still have to deactivate it for each activity? - Anyways, this still left me with a populated list of 'recent files' in quite some 'Open file' dialogs, e.g. in VS Code. I found that these dialogs are actually some Gnome/GTK/Qt dialogs instead of the KDE file picker, and the list of recently used items is being saved in '~/.local/share/recently-used.xbel'. There are various answers on the internet on how to disable this collection. - One way is to tell the programs to use the KDE file picker instead. It seems that setting the environment variable 'GTK_USE_PORTAL=1' achieves this, although according to the original post where I found this, some GTK programs may not adhere to this setting. But even if a program does accept this setting, it only does not show the recent files collection in its dialogs anymore, but the collection in '~/.local/share/recently-used.xbel' is created nevertheless. - Ostensibly, altering gnome settings, in particular gsettings set org.gnome.desktop.privacy remember-recent-files false gsettings set org.gnome.desktop.privacy remember-app-usage false gsettings set org.gnome.desktop.privacy recent-files-max-age 0 should stop the collection, but it had no effect whatsoever for me. - Another suggestion is to set settings for the gtk backends (all individually...), i.e. in '~/.config/gtk-3.0/settings.ini' and '~/.config/gtk-4.0/settings.ini' add 'gtk-recent-files-enabled=0' under [Settings]. This removed the recent files tab from the 'Open file' dialogs for me, but it did not stop the system to collect the recent files information in recently-used.xbel. - Another suggestion is to delete the file recently-used.xbel altogether and instead create a symbolic link: ln -vs /dev/null ~/.local/share/recently-used.xbel. But in my case it took about a day until this link was deleted and a new recently-used.xbel file was created and started populating again. - The gnome-control-center has a setting 'File history' under 'File history and trash', which should enable/disable the collection of recently used items in recently-used.xbel, but this setting does not work for me, it still collects items in recently-used.xbel, even with the setting turned off. It does however offer a button to erase the history and leaves a recently-used.xbel with a header but without items. But with the next opened file this list gets populated again. So none of those options worked for me out of the box. What I did was empty the file 'recently-used.xbel' with gnome-control-center (see above) and then chattr +i ~/.local/share/recently-used.xbel to make the file immutable to avoid modification or deletion by any program without root rights. This is the only option that consistently worked for me. - Many programs have their own file history features which usually can be disabled one way or another. - In Dolphin, I removed the tabs 'Recently used files' and 'Recent places' and in the Dolphin settings set it to open up at my home folder instead of the last opened place. - In the VLC Player settings at 'Interface/Main Interfaces/Qt' I unchecked 'Save the recently played objects in the menu'. - In the LibreOffice options at 'LibreOffice/Advanced/Expert settings/org.openoffice.Office.Common' I set the value of 'History' to 0, which keeps the recent files list empty. So this is where I stand now, with most of the stuff disabled, but asking myself if I missed something. And wondering why it was so complicated, because I feel that it should not be so hard to give this whole thing one single switch or maybe a few switches but all in one place in some system settings manager. I understand that if any program with its own logic keeps track of opened files, then the system cannot really stop it, but everything else should be accessible much better than what I have encountered. And why are so many options outlined above simply not doing what they say they do? Bug? My stupidity/naivety? Sorry for the long post but I wanted to demonstrate how cumbersome this really is. __TL:DR__: Why is it so hard to disable usage tracking functionality under the guise of "recent files/recent apps/recent places"?

Tor Browser conceals the user's operating system by spoofing user agents. Currently it is Mozilla/5.0 (Windows NT 10.0; rv:102.0) Gecko/20100101 Firefox/102.0. I decided to test whether the function actually does its job by visiting multiple detection websites on Tor Browser in both Windows and Linux then comparing the results. 1. GIMP Download Page(https://www.gimp.org/downloads/) 1) Tor Browser on Linux (we think your OS is Linux) 2) Tor Browser on Windows 11 (we think your OS is Microsoft Windows) 2. https://bowser-js.github.io/bowser-online/ 1) Tor Browser on Linux

Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
        
        Firefox/102.0
        browser
        	name	"Firefox"
        	version	"102.0"
        os
        	name	"Linux"
        platform
        	type	"desktop"
        engine
        	name	"Gecko"
        	version	"20100101"

2) Tor Browser on Windows 11

Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101 Firefox/102.0
        
        browser
        	name	"Firefox"
        	version	"102.0"
        os
        	name	"Windows"
        	version "NT 10.0"
        	versionName "10"
        platform
        	type	"desktop"
        engine
        	name	"Gecko"
        	version	"20100101"

3. https://useragentstring.com 1) Tor Browser on Linux

Mozilla/5.0 (Windows NT 10.0; rv:102.0) Gecko/20100101 Firefox/102.0
        Firefox 102.0
        	Mozilla	MozillaProductSlice. Claims to be a Mozilla based user agent, which is only true for Gecko browsers like Firefox and Netscape. For all other user agents it means 'Mozilla-compatible'. In modern browsers, this is only used for historical reasons. It has no real meaning anymore
        	5.0	Mozilla version
        	Windows NT 10.0		Operating System: Windows 10
        	rv:102.0	CVS Branch Tag The version of Gecko being used in the browser
        	Gecko	Gecko engine inside
        	20100101	Build Date: the date the browser was built
        	Firefox	Name : Firefox
        	102.0	Firefox version

2) Tor Browser on Windows 11

Mozilla/5.0 (Windows NT 10.0; rv:102.0) Gecko/20100101 Firefox/102.0
        Firefox 102.0
        	Mozilla	MozillaProductSlice. Claims to be a Mozilla based user agent, which is only true for Gecko browsers like Firefox and Netscape. For all other user agents it means 'Mozilla-compatible'. In modern browsers, this is only used for historical reasons. It has no real meaning anymore
        	5.0	Mozilla version
        	Windows NT 10.0		Operating System: Windows 10
        	rv:102.0	CVS Branch Tag The version of Gecko being used in the browser
        	Gecko	Gecko engine inside
        	20100101	Build Date: the date the browser was built
        	Firefox	Name : Firefox
        	102.0	Firefox version

So Tor on Windows was able to hide the fact that Windows 11 was run instead of 10. But the Linux version was unable to hide the fact that Linux was the OS. I ran all tests with no extensions except the built-in

. Also 2-b) shows an additional

; x64;

that is not present in the user agent. How are websites able to deanonymize your OS? Are there subtle differences in implementation that cannot be covered by user agent spoofing? Is there a way for Tor on Linux to blend in with Tor on Windows?

(Example screenshot taken from https://askubuntu.com/questions/1343872/ubuntu-shows-other-users-are-logged-in-whenever-i-shut-down-even-though-i-am-the) But my question is: how does the system know this? I have done sudo chmod o-r /var/run/utmp sudo chmod o-r /var/log/wtmp And I verified that my user cannot run w or who anymore: the command output is empty because of read permission denied in utmp and wtmp. So how does the system know this? Where is the information leaking?

I noticed my VM with a fairly barebones Debian 11 install periodically tries to connect to debian.map.fastlydns.net on port 80. It looks like this domain is what the DNS record of http://deb.debian.org points to (and has [for at least a year](https://old.reddit.com//r/debian/comments/kexgjr/domain_debianmapfastlydnsnet_really_debianrelated/)) , so it seems to be a legitimate request somehow related to package update tracking. But what is doing this? Debian doesn't have a MOTD showing available updates, and it usually seems to need manual running of apt update and apt upgrade to keep the system current. Could I opt-out of whatever service(s) cause these requests?