Geo-Parity

Multiple tool clusters have genuinely blurry boundaries. The residency family is the clearest problem: check_travel_residency explicitly supersedes check_schengen, check_residency, and check_tax_residency_risk, yet all four remain live, inviting misselection. Cost tools also overlap (compare_countries, compare_total_cost, compare_housing, estimate_micro_costs, calculate_ppp all return price-level figures), and the sheer 25-tool compare_* family means an agent must navigate near-synonymous verbs like compare_countries vs compare_total_cost even with excellent descriptions.

The dominant verb_noun pattern is strong and heavily used (25 compare_* tools, plus calculate_*, check_*, find_*, estimate_*), which gives the set a recognizable rhythm. However, roughly a fifth of the tools use noun-phrase or adjective-noun names that break the convention (travel_advice, passport_power, property_rules, legal_profile, health_prep, arrival_cost, required_rate, withholding_tax), and the check_* vs compare_* split is semantically fuzzy since both verbs describe information-gathering rather than clearly distinct actions.

57 tools is more than double the 25+ threshold for 'too many' and creates a heavy selection burden even for a genuinely broad domain. The 25 near-identical compare_* verbs alone are overwhelming, and whole families (four residency checkers, four find_*_visas tools) could be consolidated into parameterized tools without losing capability.

The expat/remote-worker/FIRE decision surface is covered with unusual thoroughness: entry routes, arrival costs, monthly costs, income, six tax tools, housing, healthcare, safety, connectivity, climate, work norms, banking, property, law, pensions, and retirement all have dedicated tools, plus an aggregator (country_dossier) to tie them together. Minor gaps exist—ordinary employee work visas, education quality, internet censorship, and utilities/currency controls have no tool—but none creates a dead end for the core use cases.