Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations declare readOnlyHint=false, idempotentHint=false, destructiveHint=false. Description adds rich context: returns agent:// address and key shown once, refusal semantics (already_on_bus), the force_new_account capability, and the post-signup sequence (bus_configure, doorbell, bus_invite/bus_mint).
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.