Me too. No clean solution yet. As workaround I implemented first an "extra" control in specific jobs that never should be run in PROD to block execution based on environment variable in all clusters (I don't really like much but it was effective). As a second workaround, only updating pipelines in a customized way to not include jobs from the beginning as explaining first in this topic or deleting them upon being deployed in order to not make your folder code structure too much complex, I don't really like much either xD, but what to do instead?