Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare non-read-only, non-idempotent, non-destructive, open-world. The description adds genuinely non-obvious behavior beyond them: the new secret is shown only once and the receiver must be updated, a one-time-reveal trait an agent must act on. It omits whether the previous secret is invalidated immediately or during a grace period, which matters for a rotation tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.