Knmi

The set contains several overlapping clusters: ask_pipeworx / ask_pipeworx_beta / ask_pipeworx_grounded / deep_research all route or answer questions, bet_research / polymarket_edges / polymarket_arbitrage overlap on betting signals, and entity_profile / company_facts / resolve_entity / recent_changes overlap on company data. Descriptions are unusually detailed and mitigate much of the confusion, but an agent still faces real misselection risk.

Names are almost all snake_case, but conventions are mixed: noun-first (company_facts, entity_profile, polymarket_edges, knmi_datasets), verb-first (ask_pipeworx, discover_tools, resolve_entity, validate_claim), and bare verbs (remember, forget, subscribe, unsubscribe). The pattern is readable but not predictable.

39 tools is very heavy; for a server named Knmi, only three tools are KNMI-specific while the other 36 span unrelated finance, prediction markets, AI marketing, memory, subscriptions, and meta-routing. Even as a broad Pipeworx connector, this exceeds a well-scoped tool surface and makes the set feel like several servers merged together.

For the KNMI weather-data domain, the surface stops at dataset listing, file listing, and signed URL minting; there is no tool to query or decode NetCDF/HDF5/GRIB, extract observation or forecast values, or inspect station metadata. The unrelated meta/query tools may route around this, but the KNMI-specific lifecycle has significant gaps.