Requesting some one from UI>browser>password to take a look into the issue.

Jan, I think you built this piece. Can we influence the fhe file permissions?

This is related to exporting passwords, right? In this case it was not me but Christos who built this [1]. Glancing quickly at the relevant code, it looks like this calls into base::WriteFile, which hard-codes mode 666 [2]. However, there also exists base::SetPosixFilePermissions [3], likely allowing us to change that once the file was created. Assigning to Christos for further investigation.


[1] https://crrev.com/c/782722
[2] https://cs.chromium.org/chromium/src/base/files/file_util_posix.cc?l=836&rcl=09e747504f7caf471e94acc4916864790517e40c
[3] https://cs.chromium.org/chromium/src/base/files/file_util.h?l=231&rcl=09e747504f7caf471e94acc4916864790517e40c

In  issue 409416 , it was decided that files created by Chrome should follow the users' umask.
I think it is reasonable to be more restrictive about exported passwords (then again, I think this should be the default and not an exception for password manager). I will see about a fix.

In any case, if a user shares the OS with other users, they should investigate whether their umask is appropriate for them. 

(Yes, umask 022 is correct. That's why Linux comes with it. (And not
000.) If you feel it is incorrect you can report a bug to Linux.
And/or all users were on permanent lockdown, then there wouldn't be a need
for the permissions systems at all.
Only e.g., mbox operations (and files containing passwords) are usually 077.)