It is simple. Do NOT allow the SCSP agent to communicate with the Management Server when creating the Gold Image.
Here is how CSP registration works:
1) The agent is installed on the asset
2) Once the CSP service starts for the first time, it attempts to connect to the SCSP Manager using the preconfigured IP address/FQDN that you gave it when you installed. It also uses the .ssl certificate that you pointed to.
3) When the agent contacts the manager for the first time, the Manager gives the Agent a Globally Unique Identifier (GUID) that is saved on the agent locally, and entered into the SCSP Manager database.
That GUID is the key to each agent's individuality. If you allow an agent to connect to the manager during the creation of a Gold Image that you will use later in deployment, then you will essentially give out the same GUID every time you deploy that Gold Image, causing all the CSP assets to show up as the same machine in the console.
4) You can clear an agent's GUID by forcing a re-registration. The syntax is "sisipsconfig -forcereg". Note that Permission must be given to allow the use of the SCSP tools to do this.
So, as mentioned, the key to making a Gold Image is to not allow communication to the Management Server. You can accomplish this by either blocking access to the Management Server at the network level (disconnect the Ethernet cable, use a firewall, configure a switch, etc), you can give the install a dummy IP address which you can change later on (sisipsconfig -host xxx.xxx.xxx.xxx), or you can use an incorrect .ssl certificate, which would prevent the agent from communicating with the manager (then you can replace later using sisipsconfig -certfile).