this video shows all the chipmunk joint types: youtube.com/watch
when you add a joint, there are various properties you set, mainly where the joint positions are located.
You can choose between various ways to input the positions, such as:
Layout- which will take a position in the layout.
Rect- which is a relative position to the object (for example; x=5 would be 5 pixels right of object origin, and y=10 would be 10 pixels below object origin).
Polar- which is an angle and distance from an object's origin point (so an angle of 270 point up, and distance 10 would be 10 pixels up from object's origin point).
Imagepoint- which is a current position of an imagepoint on the object (the second parameter would not get used in this case; only the first parameter which is set to an imagepoint index number).
Some joints require a different number of parameters to be set based on the type of joint it is.
You would add the action to one object, and then supply the UID of the other object it is connecting to in the parameters.
There is also a Tag property that can be set so that you can use other useful expressions after creating the joints. The tag is basically a unique name for the joint. You supply this tag in expressions to retrieve properties of the joint after it is created. So you can destroy joints with tag "uniquename" for example, or return the joint type, or the current anchor position, etc..
there's a lot of info in the behavior thread: construct.net/forum/extending-construct-2/addons-29/behavior-chipmunk-physics-82151