Skip to contentSkip to conversation

more thoughts on the ssh policy

DISCUSSION Russell Coker 2 participants 2 messages Jul 26, 2002 – Jul 27, 2002 SELinux security module

lore
Share & tools
lore
#0 lore raw

Do we need two types for the files under ~/.ssh?

With the current setup in my policy tree the sshd can read the ~/.ssh/id* files from all users, which of course could be useful in attacking other machines.

Of course if you have a bug in sshd then you can (in theory at least) cause it to execute a shell in the user's security context which runs "scp" to transfer the ~/.ssh/id* files to a host controlled by the attacker. Although such exploits may be easier said than done (maybe there's a bug in sshd that doesn't allow executing arbitary code but merely allows reading arbitary files and sending the result out the TCP socket).

What do you think?

If I am to use two types then I will have to modify ssh-keygen to use open_secure...

Signature · 3 lines
I do not get viruses because I do not use MS software.
If you use Outlook then please do not put my email address in your
address-book so that WHEN you get a virus it won't use my address in the
Quoted 1 line

From field.

--
You have received this message because you are subscribed to the selinux list. If you no longer wish to subscribe, send mail to [email protected] with the words "unsubscribe selinux" without quotes as the message.

#1 lore raw
Quoted 5 lines · Russell Coker

Do we need two types for the files under ~/.ssh?

With the current setup in my policy tree the sshd can read the ~/.ssh/id* files from all users, which of course could be useful in attacking other machines.

Sounds like a good idea.

Quoted 6 lines

Of course if you have a bug in sshd then you can (in theory at least) cause it to execute a shell in the user's security context which runs "scp" to transfer the ~/.ssh/id* files to a host controlled by the attacker. Although such exploits may be easier said than done (maybe there's a bug in sshd that doesn't allow executing arbitary code but merely allows reading arbitary files and sending the result out the TCP socket).

Also, don't forget ssh-agents socket files, eg:
/tmp/ssh-XXDe4IRT/agent.768

Ideally, only trusted user programs should be able to run ssh in a domain that has access to these sockets. That might mean any *interactive* shell, cvs, rsync, etc.

ie. it may not be sufficient just to allow ssh to access it, if an "untrusted" program can execute ssh, as this program could execute "ssh remotehost rm -rf ."

Getting all of these requirements might be difficult, so a compromise may be required.

Signature · 6 lines
Brian May <[email protected]>

--
You have received this message because you are subscribed to the selinux list.
If you no longer wish to subscribe, send mail to [email protected] with
the words "unsubscribe selinux" without quotes as the message.