Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Goes well past the readOnlyHint/openWorldHint annotations: discloses the data source (public HTML scraping), a known data-quality quirk (names derived from URLs can be shortened), what is absent from the page, and the full cost model (x402 metering, per-successful-call charge, no charge on failure).
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.