question

And here is the technical answer, in case anyone wants the details :-)

If you open up the code of any trigger, you will see that current is defined as

Object current = ownerobject(c);

The c variable is like a "magic" variable that always* points at the currently executing code. For the entry trigger of a queue for example, this is the entrytrigger node in the Queue's variables section, e.g. /Queue1>variables/entrytrigger. The ownerobject function basically strips of the 'go into the variables part' after the > and gives you the queue object back.

In a user command, you can also access the c variable, and it will point at the place where the User Command code is stored, e.g. /Tools/UserCommands/TestCommand/code. It doesn't matter whether you call the user command from a Script Window, User Event or other trigger, c will always be the user command's code node. So even if you copied the line that determines current literally to your user command, you would not get the expected result:

// c is /Tools/UserCommands/TestCommand/code
Object current = ownerobject(c);
// current is now... MAIN:/project/model !

As @steven.hamoen already answered, the way around this is by passing the current from the object trigger as an argument into your usercommand,

// In the on entry trigger
Object item = param(1);
Object current = ownerobject(c);
int port = param(2);
TestCommand(current);

and then getting it out in the user command by

// In the user command
Object current = param(1);

* Actually there are some subtleties, such as when executing event functions.