Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=false, destructiveHint=true, idempotentHint=true, and openWorldHint=false, so the safety profile is largely carried by structured data. The description adds the schema-lookup prerequisite but does not disclose a key behavior for a destructive mutation: whether omitted fields are preserved or cleared, or what happens to fields not supplied. That is a meaningful gap given destructiveHint=true.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.