Scored across 33 tools
Every tool targets a distinct resource and analytical view: events.list, events.observed_schema, and events.property_values cover counts, schema drift, and value space respectively, while traffic, revenue, errors, funnels, cohorts, and search tools each have clearly separated purposes. Cross-references and 'Pairs with' notes reinforce tool selection.
Names consistently use lowercase dotted domain prefixes (events.*, revenue.*, errors.*) with snake_case, making categories predictable. The second segment mixes verbs (list, create, delete, compare) with noun views (overview, breakdown, observed_schema, journey), so it is not a uniform verb_noun pattern but remains coherent.
At 33 tools the surface exceeds the 25+ threshold and is heavy for an agent to hold in context, even though the analytics domain is broad. The clear namespacing prevents chaos, but the sheer number of tools makes the server feel over-packed for an MCP endpoint.
The tool set covers the full analytics lifecycle: traffic and event measurement, revenue/summary/retention, funnels, cohorts, errors, alerts, search-console integration, user journeys, and session/page engagement. Saved objects support create/list/delete with documented delete-and-recreate workflows for edits, so there are no dead ends for the stated scope.
