Re: "sniffer" category

[David Fifield <david () bamsoftware com>]

On Wed, Nov 09, 2011 at 12:27:20AM +0100, Patrik Karlsson wrote:
> On Wed, Nov 9, 2011 at 12:01 AM, David Fifield <david () bamsoftware com>wrote:
>
> > On Tue, Nov 08, 2011 at 09:43:30PM +0000, Luis MartinGarcia. wrote:
> > > On 11/08/2011 08:27 PM, David Fifield wrote:
> > > > On Tue, Nov 08, 2011 at 05:23:52PM +0100, Patrik Karlsson wrote:
> > > > > I'll check the broadcast-listener script for this as well. In regards
> > to
> > > > > these sniffing scripts I would like to create the "sniffer" category
> > and
> > > > > place them in there, rather than in the broadcast category as we've
> > > > > discussed earlier.
> > > > >
> > > > > I guess that the new category needs to be documented somewhere in
> > addition
> > > > > to changing the category in the scripts? Where would that place be,
> > and is
> > > > > "sniffer" the category name to go with?
> > > > Is "sniffer" really what we want to express? It seems to me what people
> > > > want is a category for "scripts that run on the whole network with a
> > > > fixed delay that I don't care about when I'm just scanning a few
> > hosts."
> > > > I think that people use "broadcast" with that meaning now, mostly in
> > the
> > > > form "and not broadcast". So "broadcast" might not be the right name
> > for
> > > > the category, but breaking out a separate "sniffer" is just going to
> > > > make people change to "and not broadcast and not sniffer".
> > >
> > > I'm not entirely familiar with the current status of NSE scripts but, in
> > > my opinion, it'd be good idea to group all those scripts that gather
> > > information passively by capturing incoming packets. However, I'd name
> > > the category "passive", not "sniffer". I think "passive" scripts can be
> > > quite useful in penetration testing when one does not want to inject
> > > packets into the network. We could perhaps define the "active" alias as
> > > "all and not passive". Does this make sense?
> >
> > "passive" is not a good name. Some of these scripts do in fact send
> > traffic (broadcast-dns-service-discovery is an example). What makes
> > these scripts different is that they do not target the hosts you give on
> > the command line. When I scan scanme.nmap.org with --script=safe, I
> > don't want a bunch of scripts telling me about things on my local
> > network. I really think that's what this is about, not
> > unicast/broadcast, sniffer, or active/passive. If the "broadcast" name
> > really bothers people, can we think of a name that reflects what this
> > category is actually used for?
>
> Well, I think the broadcast name is good for the category of scripts that
> actually do send broadcast and multicast traffic.
> However, there are at least two broadcast-listener and targets-sniffer that
> are passive as they don't send any data.
> These are the ones I was thinking moving to a new category, as they differ
> from the rest of the broadcast scripts.
> But, maybe we should wait until we get more of them until we do, I don't
> know ....

I don't have a problem with giving those scripts better category names.
But there needs to be a new category that includes all the scripts that
are currently in "broadcast". A "sniffer" category may be nice to have,
but "and not broadcast" serves a very real use case that we can't throw
away.

A short-term solution would be to add the "sniffer" category to those
scripts, but not take them out of "broadcast". I'm not quite convinced
of the need for "sniffer" though. "broadcast" is not a good name for
what it's used for, even if what it's used for is useful. Something
needs to fill the void if it stops meaning what it does now. I'm fine
with separate or overlapping "broadcast" and "sniffer" categories, as
long as there is a new category that means "broadcast or sniffer" to
make them easy to turn off.

David Fifield
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/