Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare a non-read-only, non-idempotent, destructive write. The description adds genuinely new context beyond them: the reload is queued rather than executed synchronously, and completion remains unknown. That is exactly the kind of async/fire-and-forget caveat an agent needs. It still omits whether the zone is interrupted and whether config validation failures abort the queue.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.