Nameidea

The two tools occupy clearly separate roles: search_surnames finds records by filter, and get_surname_record retrieves a specific record by the stable ID that search returns. Their descriptions even cross-reference each other and delineate boundaries (get's not_found only applies to the 3,000-record subset). An agent cannot reasonably confuse them.

Both names follow a clean snake_case verb_noun pattern (search_surnames, get_surname_record). Plural/singular usage is logical (search a collection, get a single record) and consistent throughout.

Two tools is on the thin side, but the server explicitly scopes itself to surname lookup over a narrow dataset so search+get may be intentional. Still, for the apparent surface (Census records plus curated collections) a single search and a single fetch leaves little room for related operations.

Search plus retrieve-by-ID covers the core read-only lifecycle for a lookup API, and the descriptions honestly note that meanings and availability checks are out of scope. Minor gaps remain: no way to browse/list the curated collections or fetch multiple records at once, which agents must work around.