Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly, idempotent, non-destructive, closed-world, but the description adds real behavior: results are ranked, capped by `limit`, and include kind/id/name/snippet. It also discloses two non-obvious traits the annotations cannot express — archived items are silently excluded, and emails/phones are returned as '[redacted]'. Only ranking criteria and pagination beyond `limit` are left unstated.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.