Re: ncat suggestions...

[David Fifield <david () bamsoftware com>]

On Fri, Nov 06, 2009 at 05:15:29PM +0000, MALTE SIMON wrote:
> Hi I suggest the following changes for ncat plese comment them ;-)

Thank you for your suggestions. I have made some quick comments. It
would help if you could describe what the changes would be used for, as
task someone would want to accomplish that is made easier with each of
the changes.

> 1. Make the following aliases:
> -z = --send-only --recv-only 

What is this for?

> -i = -d

There are already separate -i and -d options. -i is the idle timeout and
-d is the delay between sends.

> -L = -l -k

This one is reasonable, but our -l -k doesn't work exactly like -L works
on some Windows Netcat. (As I understand; I haven't tried it.) -L
listens for a new connection after the first connection has finished,
but -l -k keeps listening up to the connection limit while connections
are ongoing; it doesn't just respawn itself. You can connect with two
clients at the same time with ncat -l -k.

> -V = -version

> -so  = --send-only
> -ro = --recv-only

These would break the rules for option parsing. In -so, "o" looks like
the argument to -s, so you get the error "Could not resolve source
address o". There is no -r option, but -ro is parsed like -r -o. It
would be possible to hack in special cases for these two but I don't
like it.

> readd -r for using a random source / listen port.

The operating system already chooses a source port for you when you make
a connection without -p. What is the value of using a random listen
port?

> -p for choosing a port to connect / listen 

We already have -p for choosing a source port. I like our way of
specifying the host and port better than the Netcats that use -p.

> ncat should understand commands like ncat 127.0.0.1:23 --> connect 127.0.0.1 Port 23
> ncat -l 127.0.0.1:23 --> listen on interface 127.0.0.1 Port 23

I think this is a good idea.

> will the a -D to run as deamon on win systems? SBD got this maybe ther code can help
> Is ther a option (-r X) to respawn X secents after a connect (-r0 = -l -k)

No. This is probably better done with a shell one-liner. -r 0 would not
be exactly like -l -k, because as I said above -l -k doesn't respawn. It
keeps listening for new connections while connections are ongoing.

> is ther a option to toggle monitoring (snooping (-M) ) on/off (used with the -e / -c option)?

Please explain what you mean by monitoring. Logging of input and output
of subprocesses? That doesn't work but is planned to be added.

> add a custom prefix (-P ) to all outbound data mostly useful for "chat mode" like nickname

This is possible, but not trivial because you have to define what
separates each chunk of sent data. Doing it by lines would be good, but
we don't have the code in place to allow that yet. Breaking sends into
lines is also desired for the -d option.

> add a option  to make ncat  run only X instances of itself. Maybe command like -1 (for one) and -1 15 (for 15) dont' 
> know

Use -m (--max-conns) for this.

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