so your pidgin.exe starts on your local machine (127.0.0.1) and you
are able to use it. Did you configure it to connect to HotFuzz (which
is listening on 127.0.0.1:5298)? Or does it try to discover the port
through bonjour and connects to 192.168.1.5 (which would skip
HotFuzz)?
Martin
Client Proxy IP :127.0.0.1 port 8080 Program : \directory\pidgin.exe |
symbols: c:\symbols Agent IP : 127.0.0.1 port 9001 ----- Server Target IP 192.168.1.5 port 5298 (bonjour port) program: \directory\pidgin.exe symbols: c:\symbols Agent IP: 127.0.0.1 port 9002 |
-----+ |
I chose bonjour because it is serverless. 2 machines on a same network can communicate. The 2 vms i have are on a vlan and can chat.So does it seem that pidgin is not a good candidate for hotfuzz?
What would be a port that hotfuzz owns?On Mon, May 16, 2011 at 2:39 PM, Martin Žember <zem...@gmail.com> wrote:
Wireshark shows what it actually does...
The first phase - multicasting on port 5353 - serves for exchanging the machine addresses. When they know about each other, there is the other phase, usually using ports 5298... The exchanging phase establishes communication (without HotFuzz). It has to be clear what is intended to fuzz. In case of the later phase fuzzing, pidgin must be forced to communicate through a given port (belonging to HotFuzz). It turns out to be a rather difficult than easy scenario... How do the nodes find each other if they are not on the same network?
On Mon, May 16, 2011 at 8:27 PM, Joel Fernandez <joelfer...@gmail.com> wrote:ok. thanks. i'm trying to fuzz any protocols that pidgin uses. bonjour was the easiest since it is serverless and i could do it in an enclosed vmlab.On Mon, May 16, 2011 at 2:21 PM, Martin Žember <zem...@gmail.com> wrote:
Yes, we are trying to find a solution...
Beware that the HTTP proxy communication is probably not what you want to fuzz. Is it the Bonjour protocol you want to fuzz?
I've tried to reproduce it... If you setup the Bonjour port from 5298 to something else, the mDNS protocol announces the changed port, which circumvents the proxy at the end...
Will think about it more...On Mon, May 16, 2011 at 8:09 PM, Joel Fernandez <joelfer...@gmail.com> wrote:
any idea what i may be doing wrong?
Dusan,
I am also working with Hotfuzz and have done fuzzing targetting Pidgin and ichat.
I have a question about this particular configuration.
The setup is as follows
1. WinXP Machine 192.168.15.100 - Pigin Client with Bonjour Protocol - HotFuzz installed
2. WinXP Machine 192.168.15.101 - Pigin CLient with Bounjour Protocol - NO HotFuzz
To target the Pidgin application on this test, the client is the .100 machine. However, given that there is no actual server only a communication directly to the host, the .101 machine will act as the server.
The agent running on the .100 is installed as part of the Hotfuzz application.
The agent running on the .101 machine, how would this be installed? Should the HotFuzz application be completely installed and only the agent section of the screen be started?
I tell you this because it might help Joel with his configuration. Now, the Bonjour protocol is based on mDNS queries to find other users on the network. To target Pidgin, should the configuration of the xml files be changed to have target mDNS?
Do you have examples of configurations that can be used to target Instant Messaging protocols?
thanks.
Well, it is still possible to fuzz the mDNS exchange. Just look at the communication in Wireshark to see the the right broadcast IP address and the port to fill into HotFuzz (5353, actually).
The other phase is also possible, it just needs some update in the code to change the port in those mDNS packets.
On Mon, May 16, 2011 at 8:43 PM, Joel Fernandez <joelfer...@gmail.com> wrote:
I chose bonjour because it is serverless. 2 machines on a same network can communicate. The 2 vms i have are on a vlan and can chat.So does it seem that pidgin is not a good candidate for hotfuzz?
What would be a port that hotfuzz owns?
On Mon, May 16, 2011 at 2:39 PM, Martin Žember <zem...@gmail.com> wrote:
Wireshark shows what it actually does...
The first phase - multicasting on port 5353 - serves for exchanging the machine addresses. When they know about each other, there is the other phase, usually using ports 5298... The exchanging phase establishes communication (without HotFuzz). It has to be clear what is intended to fuzz. In case of the later phase fuzzing, pidgin must be forced to communicate through a given port (belonging to HotFuzz). It turns out to be a rather difficult than easy scenario... How do the nodes find each other if they are not on the same network?
On Mon, May 16, 2011 at 8:27 PM, Joel Fernandez <joelfer...@gmail.com> wrote:
ok. thanks. i'm trying to fuzz any protocols that pidgin uses. bonjour was the easiest since it is serverless and i could do it in an enclosed vmlab.
On Mon, May 16, 2011 at 2:21 PM, Martin Žember <zem...@gmail.com> wrote:
Yes, we are trying to find a solution...
Beware that the HTTP proxy communication is probably not what you want to fuzz. Is it the Bonjour protocol you want to fuzz?
I've tried to reproduce it... If you setup the Bonjour port from 5298 to something else, the mDNS protocol announces the changed port, which circumvents the proxy at the end...
Will think about it more...
On Mon, May 16, 2011 at 8:09 PM, Joel Fernandez <joelfer...@gmail.com> wrote:
any idea what i may be doing wrong?
I am still working on this (didn't want you to think i gave up)