An async plugin step that keeps its completed system jobs

Asynchronous plugin step never purges completed jobs.

Successful async runs pile up in the system-job table without bound, slowing the async queue and consuming storage.

Check the delete-on-success setting on each asynchronous step. Off means every successful run is retained forever. The knock-on effect shows up as a large system job table and a slower async queue long before anyone connects it to the setting.

Turn on delete-on-success so completed jobs purge while failures stay for troubleshooting.

This one is environment health, so it carries a plain label and no control mapping. Presenting an operational gap as a security finding would make the real security findings harder to trust, so we keep the two apart.

• A plugin step that re-triggers itself on every save
• Two plugin steps at the same execution order writing one column
• A plugin registered twice on the same message
• An update plugin step with no filtering attributes