Scored across 23 tools
Most tools target clearly distinct resources (news, calendars, localities, meshes, SIDRA, names, countries). The main potential confusion is between the low-level ibge_sidra and the many wrappers (ibge_censo, ibge_indicadores, ibge_cidades, ibge_comparar), though the descriptions explicitly cross-reference and delineate when to use each.
The vast majority follow a consistent ibge_<resource> pattern with snake_case, e.g. ibge_municipios, ibge_sidra, ibge_malhas_tema. The two exceptions are the bare 'search' and 'fetch' tools, which break the prefix convention for a separate Deep Research contract.
At 23 tools, the set feels heavy for a single server, even though the IBGE domain is broad. The number is justified by the range of distinct data services, but it crosses the 16-25 borderline that starts to feel unwieldy.
The tool surface covers an impressively complete read-only IBGE workflow: discovery (search, pesquisas, sidra_tabelas), metadata (sidra_metadados), data queries (sidra, censo, indicadores, comparar, cidades), geography (estados, municipios, localidade, geocodigo, malhas), and auxiliary domains (nomes, paises, cnae, datasaude, noticias, calendario). No obvious dead ends remain for the stated statistical-data purpose.