Besides trusting that the limitation statement would’ve been a welcome addition to the documentation, I would also vote in favour of this feature with the ability to set the initial docker image to be used as a replica of the exec runner environment. Something like:
drone exec --image=node/13-alpine
I’m currently using an exec runner as it’s the only one running on docker containers without privileged access. i.e ACI
With that in mind, it becomes easier to isolate a docker image with all I need and install the runner in it. Since it’s already been ran from an image, using that same image as a base of the local execution would be great.