Related Servers

Most tools target clearly distinct operations: run/ranking/describe for routing, fetch_result for paging, report_outcome/report_narrative for feedback, and declare/update/remove for custom tools. A few clusters overlap, especially login/use_token/onboard/connect_info around auth and key handling, and my_preferred_tools/prefer_tool versus get_preferences/set_preferences around the word 'preference'. The descriptions usually disambiguate, but an agent could still misselect within those clusters.

All names are snake_case, but the verb/noun convention is mixed: some are clear verb_noun forms (fetch_result, set_preferences), while others are bare nouns or verbs (console, setup, run, ranking, login, describe). The my_* prefix is also used inconsistently across my_tools, my_preferred_tools, and declare_my_tool. It remains readable, but the pattern is not fully predictable.

At 22 tools, the surface is heavy for a routing server and sits in the borderline 16-25 range. Many tools are justified by the platform's scope (auth, setup, routing, ranking, reporting, preferences, custom tools, result paging), but the count still feels large and increases selection burden. It is not extreme, but it is more than a tightly scoped set.

The set covers the core lifecycle well: session/auth, setup, routing and inspection, large-result paging, outcome reporting, custom-tool CRUD, preference/facet management, console views, and connect steps. Minor gaps remain, such as no explicit logout/session reset or a plain catalog-wide tool listing, but agents can work around these via existing tools.