Currently as long as the script is pushed succesfully to the managed device, the server considers this a successful run. It would be very useful if the server was able to determine when a script exited with errors so we can see which particular machines failed to execute correctly so we can perform manual remediation.