Scored across 55 tools
The set contains a large cluster of overlapping search/intelligence tools (search_graph, get_evidence, get_domain_intelligence, get_specialist_intelligence, get_expert_intelligence, get_report_intelligence, search_statistics, search_insights, get_intelligence_dossier) whose boundaries rely on subtle, description-embedded distinctions an agent can easily misjudge. More damning, verify_market_claim and verify_claim have essentially identical descriptions, guaranteeing misselection. Several tools do the same thing under different names.
Almost everything follows a snake_case verb_noun pattern (get_*, search_*, list_*, draft_*, consult_*, submit_*, check_*, request_*), which is predictable and readable. A few outliers like brand_tracker, brainstorm_topic, expert_onboarding_research, and deep_research_topic drop the verb, but the convention is largely consistent.
55 tools is far beyond a well-scoped surface and well past the 25+ 'too many' threshold, signaling bloat. The count reflects many overlapping intelligence/search variants plus onboarding, deliverables, and account tools bundled into one server, making the surface hard to navigate.
The surface is broad and covers account management, graph discovery, search, earnings intelligence, expert onboarding, consultation, deliverables, feedback, and scheduled reporting with few obvious dead ends. Minor gaps exist but the domain lifecycle is largely covered.
