Related Servers

The bulk of the surface follows a clean resource+action split (list/get/create/update/delete per entity), and the domain-specific tools have clearly differentiated purposes (audit_documents vs verify_einvoice vs preview_totals, send_einvoice vs bulk_send_einvoice, mark_paid vs bank_reconcile). A few pairs still blur: schedule_run_due vs schedule_run_now, duplicate_document vs bulk_duplicate, and the catch-all api_request overlapping every specialized tool. These are resolvable from descriptions but add selection risk at 123 tools.

There is a dominant, predictable verb_noun snake_case convention across all CRUD families (list_clients, create_issued_document, update_vat_type, delete_receipt), and domain clusters use consistent prefixes (bank_*, schedule_*, bulk_*, tax_profile_*). Deviations exist but are internally coherent: noun-first report names (revenue_summary, client_statement, cashflow_forecast), noun_verb order for tax_profile_get/set, and preposition-like names (client_from_vat, quadro_lm, lookup). Readable, but not a single uniform pattern.

123 tools is far beyond a comfortable agent-facing surface, even for a broad invoicing/accounting platform; the CRUD quintets for a dozen resources alone account for most of it. Several tools could be consolidated (e.g. generic list/get/create/update/delete parameterized by resource, or the many report tools grouped). Each tool has a real function, so it is not a 1, but the count materially burdens discovery and selection.

Coverage is unusually deep: full lifecycle for issued/received/archive documents, clients, suppliers, products, receipts, F24, cashbook, payment accounts/methods and VAT types, plus e-invoicing, reconciliation, tax estimation and reporting workflows with no obvious dead ends. Minor gaps remain for secondary resources (revenue/cost centers, payment terms, document templates, webhooks) that are only reachable via lookup or the api_request escape hatch. Strong but not exhaustive.