Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare the safety profile (readOnlyHint=false, idempotentHint=true, destructiveHint=false), so the description carries a lighter burden. It usefully identifies the staging-area target, which is the core behavioral effect and is not encoded in any structured field, but it omits what forms the 'files' argument accepts (paths, globs, directories) and doesn't clarify that nothing is committed.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.