Performs a bulk action on a set of jobs. The job set is determined by an explicit job_uid array and/or filter_rules (v3) / legacy filter.* fields, combined additively.

assign_to_route and auto_assign_territory are asynchronous: they queue a worker and respond with an acknowledgement before the work completes. delete responds before an async cascade-cleanup worker runs. send_email and send_sms respond before their per-job send loop finishes. update_schedule and clear_schedule are fully synchronous (one save per job) and are blocked with 403 if the company has the Appointments feature enabled. update_status applies to all matched jobs except those skipped for exceeding the job-status-history limit or failing a status-dependency check (when mandate_dependency_check is set) — those are reported back in data, the rest still succeed (partial success, not an all-or-nothing failure).