Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare idempotentHint=true and openWorldHint=true, so the agent knows re-running is safe and the tool reaches beyond a closed set. The description adds that it both regenerates CSS and flushes the object cache, which clarifies side effects on cached data. It does not mention permissions, performance cost, or any downtime, so it adds only moderate context above the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.