Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond the annotations (read-only, idempotent, non-destructive), the description discloses exclusions ('soft-deleted messages and pending/declined-invite groups'), strict timestamp filtering, ordering, and clarifies response fields (`count`, `has_more`, deprecated `total`). This adds significant behavioral context not available in annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.