Check the latest deploy run before calling a pipeline done

Before answering 'does it all work now', run gh run list and look at the most recent deploy workflow run, not just CI. In this repo CI passed on the head commit, but the workflow_run-triggered Databricks bundle deploy failed: the serverless notebook task could not pip install kagglehub==1.0.3 (PipError against the serverless package constraints files). Successful dev runs and green CI can coexist with a broken production run. To pull the decisive error line quickly, use gh run view <id> --log-failed | grep -iE 'error|fail'.