Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With annotations carrying essentially no safety profile, the description carries the full behavioral burden. It discloses that the secret and first data key are shown exactly once, that the account gets a 1,000-credit grant, that duplicate addresses return 409 email_exists with next_action login_to_link, and how the configure block relates to future fetch_data calls. This is rich, non-obvious behavioral context.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.