Pitchprotocol

Multiple tools have overlapping or deliberately duplicated purposes: ask_pipeworx_beta explicitly mirrors ask_pipeworx, ask_pipeworx_grounded and deep_research blur into the same routing space, and the prediction-market cluster (polymarket_edges, polymarket_arbitrage, polymarket_edge_tracker, polymarket_fill_risk, polymarket_kalshi_spread, bet_research) presents several near-adjacent 'find an edge' tools. An agent would struggle to reliably pick the right one without deep description parsing.

Names are uniformly snake_case and follow recognizable clusters (ask_pipeworx*, polymarket_*, pitchprotocol_*, resolution_*, pipeworx_*). There is some mixing of verb-first (compare_entities, validate_claim) and noun-first (entity_profile, company_facts) forms, but the overall pattern is predictable and readable.

38 tools is well above the 25+ threshold and the surface feels like a grab-bag of unrelated domains: a data-lookup router, prediction-market analysis, AI-visibility scanning, npm dependency checks, memory storage, subscriptions, and the nominal Pitch Protocol submission flow. The count is not justified by a single coherent scope.

Within its broad actual scope (structured data lookup, prediction-market analysis, memory, subscriptions) coverage is reasonably complete, including grounded answering, resolution auditing, and subscription lifecycle. However, the server's stated name 'Pitchprotocol' is served by only two tools (schema + status) with no create/submit tool, and the memory/subscription features feel bolted on rather than integral.