Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint, idempotentHint and openWorldHint=false, so safety is covered; the description's contribution is disclosing the shape of the returned data (per-fact evidence status, source ids, checked dates, missing information), which is genuinely useful since there is no output schema. It still omits likely-failure behavior (unknown or malformed slug) and any notion that no filtering/pagination applies.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.