As I understand it right now, if I specify an agent, it creates a node, runs checkout scm, and uses the same node/workspace for all of the stages? And if I specify agent: none, I have to create my own node in each stage, and re-run checkout scm for each stage?

What I want to do is specify "Additional Behaviors" on the checkout step (currently using the workaround described here), but achieve the benefits of running it once on the top-level agent and re-using across all stages.

Is there any way to achieve this? Either by specifying Additional Behaviors on the top-level pipeline, or allowing a user to override the implicit checkout scm that happens when agent is set?

Thanks!