Output ZDepth from Point Render
If you really want to get the zdepth to work in the current version, you could try having a second Point render node in facet render mode, world position mode and depth buffer turned on. This will give you world position pass of your point cloud rendered through the camera (if it doesn't crash!). Then multiply the inverse camera matrix by the world position pass to normalize the position data to the camera, then the blue channel becomes the depth channel! It doesnt always work but may be fun to play with. I attached an example script below: https://drive.google.com/open?id=12zrLiD27R-8WQCD7tsQnC-nmQ4YtLLWX
As rgray point's out you do need a second point render node to output depth.
Sadly blinkscript does only allow you to output 4 channals, and as Red Green Blue and Alpha is needed, I need to run the whole render a 2nd time just to get depth out, but that is an option in the 1.1 update (just around the corner).