agent-eligibility

Most tools have clearly distinct purposes: check_sanctions and check_vat target different registers, evaluate_eligibility scores agent/task pairs, and the two metadata tools are unambiguous. The only mild overlap is check_counterparty, which composes check_sanctions and check_vat, so an agent could momentarily wonder whether to call the composite or the components—though the description explains the relationship.

Every tool follows a consistent snake_case verb_noun pattern (check_*, evaluate_*, get_*, list_*). The verbs accurately reflect the actions and there is no mixing of conventions.

Six tools is a well-scoped set for a counterparty/eligibility checking service. Each tool earns its place: two atomic checks, one composite, one scorer, and two metadata/auxiliary tools.

The surface covers the core lifecycle: atomic sanctions and VAT checks, a composed verdict, an eligibility scorer, and the published scoring rules. Minor gaps exist—no batch evaluation, no results history/retrieval, and no way to enumerate supported tasks—but core workflows are covered.