Usgs Mrdata

Many tools overlap heavily as universal routers (ask_pipeworx, ask_pipeworx_beta, ask_pipeworx_grounded, deep_research, validate_claim) and as company-data entry points (entity_profile, company_facts, compare_entities, resolve_entity, recent_changes). The very long descriptions do explicitly state when to prefer each, so boundaries are mostly inferable, but misselection among the router family remains likely.

All names are snake_case, which aids readability, but the set mixes verb_noun (resolve_entity, validate_claim, search_within), noun phrases (company_facts, entity_profile, bet_research, polymarket_edges), and prefixed families (mrdata_*, pipeworx_*, polymarket_*, ask_pipeworx_*). No single predictable pattern holds across the 39 tools.

39 tools is heavy for the server's apparent USGS MRData purpose, and only 3 tools (mrdata_search_deposits, mrdata_deposit, mrdata_commodities) actually serve that domain. The rest are a broad multi-domain gateway spanning prediction markets, memory, subscriptions, AI visibility, and routing, making the count excessive relative to any single stated scope.

Read-only coverage of USGS MRDS is reasonable (search, detail, commodity lookup), and meta-areas like memory and subscriptions have full create/read/delete cycles. However the surface has notable gaps: no direct market-fetch tool despite resolution_audit/resolution_diff referencing polymarket_market and kalshi_market, and most data domains are reachable only indirectly via ask_pipeworx rather than exposed tools.