On Thu Aug 19 2004 at 14:44, Marc Haber wrote:

> The issue is not only "dangerous for the scanning server", but also
> dangerous for the scanned systems. Imagine a trojaned plugin not only
> detecting a remote exploit, but also actively exploting it to install
> a backdoor.

Of course, anything is possible, but IMHO the probability of this
happening is quite low. On the other hand, a simple plugin can crash a
fragile service; although we try to lower this risk, this already
happened and this will happen again in the future, unfortunately. I
don't think that more audit will change anything.

> the scanned systems to execute the local plugins that might have access
> to a Windows administrator account or to an ssh key that is allowed to
> execute arbitrary commands on a target system.

SSH key are stored in a protection section of the KB that can only be
read by _trusted_ plugins.
Concerning the Windows administrative rights, the issue is more tricky
because we'd have to "certify" many (125) plugins. We're thinking of
it...

> I don not intend to express any distrust for Renaud

Said in another way, if cvs.nessus.org is hacked, it will simpler to
put a Trojan horse into a C file than a NASL script.

> explain why there might be situations where it is desired to have the
> Debian maintainer as another pair of eyes looking at the code.

There are already many eyes here, on this list or nessus-devel, which
are much more efficient than eyes from distro maintainers. Are you
sure you need one more pair and lose a couple of weeks? 
_______________________________________________
Nessus mailing list
[EMAIL PROTECTED]
http://mail.nessus.org/mailman/listinfo/nessus

Reply via email to