Given that you are using Wireshark, your port numbers aren't automatically resolved to an application name, so you will need to do a little more to refine the information you are looking for. Every application using TCP/IP to communicate across a network will be using ports, so that the network stack knows where to deliver segments to (I like to call it an application address).
Clients connecting to a server application on a specific port will be dynamically allocated a port number from a dynamic range. So you first need to find out what TCP/UDP connections your application has open:
netstat -b
at the command line will give you a list of connections with the name of the executable that created the connection. Each executable has one or more connections listed as 127.0.0.1:xxxxx, where X is the local port number for the connection.
Now in wireshark, you need to tell it to display packets that originated from or are destined to that port by using one or more of the these filters:
tcp.port == xxxxx or udp.port == xxxxx
Add an additional or tcp.port == xxxxx for each connection you want to show.
This will allow you to see all the traffic for the connections your application has open and Wireshark will not include just raw TCP/UDP segments but it will include the various application layer protocols (eg. HTTP) that used those port numbers too.
If your application appears to be communicating with just one server, you could just use the IP address of that server to filter by:
ip.addr == x.x.x.x
Best Answer
First off, Wireshark doesn't have explicit Skype support yet so you won't get pretty decodes. If you're wiling to compile from source, you can build it in.
Second, Skype encrypts everything so there isn't much to look at once you get the packets.
Third, it doesn't use a predictable port unless egress to the network is restricted to only allow TCP/80 or TCP/443, so you'll have to find them in the swarm of normal traffic. But that should be easy if you've properly isolated your network connections.