Stooq

The tool set has several distinct clusters (data queries, Polymarket betting, memory, subscriptions), but there is notable overlap within clusters: ask_pipeworx, ask_pipeworx_beta, ask_pipeworx_grounded, deep_research, and validate_claim all route questions to the same underlying data sources with different modes, which could cause confusion about which to select. Additionally, ai_visibility_check and scan_competitor_ai_presence are near-duplicates (single vs. multi-entity), as are remember/recall/forget and entity_profile/recent_changes.

Naming is largely snake_case and descriptive, but there is no consistent verb_noun pattern: some tools are verbs (ask_pipeworx, compare_entities, resolve_entity), some are imperative actions (subscribe, unsubscribe, remember, forget), and the polymarket_* series uses a different prefix convention than the pipeworx_* series. Deep_research, entity_profile, and search_within break the pattern of prefix+noun by using a descriptive phrase first.

31 tools is on the heavy side for a single server. While the server is a broad data-research platform spanning many domains (financials, drugs, economics, prediction markets), 31 feels like it could be consolidated—several tools are variants of the same core capability (ask_pipeworx vs. ask_pipeworx_beta vs. ask_pipeworx_grounded, polymarket_edges vs. polymarket_edge_tracker). The count is within a plausible range for such a platform but at the upper edge where discoverability and selection cost become concerns.

The tool surface is comprehensive for the stated purpose: querying structured data sources, researching companies, comparing entities, verifying claims, having prediction-markets tooling (arbitrage, edges, fill risk, cross-venue spread, edge persistence), memory (remember/recall/forget), subscriptions (subscribe/unsubscribe/recent_alerts/list_subscriptions), and meta-tooling (discover_tools, suggest_questions, pipeworx_feedback, pipeworx_trending). Minor gaps: no tool for direct raw data export or visualization, no explicit file/download tool, and the patents data path is noted as soft-failing (USPTO sunset), but these are peripheral to the core data-research mission.