No user-submitted related servers found.
Scored across 33 tools
Most tools target distinct resource+action pairs, and the detailed descriptions make even similar pairs (move vs set_pose as relative vs absolute, plan_pieces vs plan_status) easy to separate. A few pairs could trip up an agent though: plan_load_code vs plan_load (site string vs saved file), hub_import vs plan_load_code (both replace the plan from an external source), and browser_push vs browser_readback.
Snake_case is used throughout and the plan_, browser_, and hub_ prefixes create strong group predictability. However, the word after the prefix alternates between noun (plan_status, plan_code, browser_readback) and verb (plan_save, plan_load, browser_push), and the bare tools mix verb-first (list_buildables, set_pose), noun-verb (data_refresh, wall_run), and imperative (undo, stack) styles.
33 tools is heavy and exceeds the typical comfortable band, but the server genuinely spans several sub-domains: buildables catalog, plan lifecycle, geometry editing, persistence, hub sharing, and browser automation. Some fragmentation exists (wall_run and ring could be one pattern generator; the seven browser_* tools could be consolidated into fewer parameterized tools), so the count earns a middling score rather than a poor one.
The tool surface covers the full base-planning workflow with no dead ends: rules and buildable data, plan creation/query/serialization, every placement and transformation primitive, validation (plan_status, perimeter_check), persistence (plan_save/load/list), sharing (plan_code, hub_import/hub_save), and full browser interactivity including readback and screenshots. Every save has a corresponding load and every import has an export.