Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond annotations (openWorldHint, idempotentHint), the description reveals that the broadcast is 'one-shot', spam-checked, automatically adds an unsubscribe footer, and fails if no subscribers exist. This adds valuable behavioral context about safety and preconditions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.