> On Apr 19, 2025, at 3:20 PM, GreyFedora via Tails-dev <tails-dev@???> wrote:
>
> How come no tools included in Tails to remove PII from database files or data masking software to protect confidential sources in whistleblowing or leaking?
Hi! I'm not a Tails maintainer, but I can take a stab at answering this question
based on observations of this mailing list. A real Tails maintainer can give you
a more definitive answer.
1. Generally the Tails maintainers want to only add software that is *already*
packaged in Debian. Thankfully, metadata-cleaner meets this requirement.
2. That software must be maintained. That's a problem. On 2025-01-23 in
message "Metadata Cleaner is looking for a new co-maintainer" it was noted that
"Unfortunately, the future of Metadata Cleaner is compromised: it has
no active maintainer anymore."
3. Generally the Tails maintainers are trying to strike a balance. They want to provide
enough 'general use" functionality so you don't *have* to install anything else, but each
package they add potentially adds new attacks, so they don't want to add *too* much.
Details below. I hope that helps.
--- David A. Wheeler
=== Details ===
Currently you can get this functionality by installing the package
metadata-cleaner, which is a GTK GUI frontend for mat2.
I've never used it, and I have no idea how good it is. A few comments here:
https://www.omglinux.com/metadata-cleaner-app-for-linux/
I suggest you investigate it, examining its results, and report back on
how well it works. It'd be great if you could find or create a test suite
to see how well it really *does* remove metadata. That would address point 1.
I don't know if anyone has stepped up to implement point 2. That
matters obviously :-).
You could also investigate how much more code it adds & whether or not
this is widely-needed functionality, meeting point 3. I could see this software
meeting the bar for point 3 ("it is widely useful for intended users").
Reporters would often need this in particular to justify claims while
shielding sources.
I could *especially* see an argument for it if the
Tails developers take additional steps to harden the software
(e.g., implement bubblewrap on some of the lower-level functions and/or
recompile some C code with hardened options like those identified by
OpenSSF).
