Scored across 36 tools
Each tool targets a distinct resource and operation, with clear separation between similar tools like get_component vs inspect (single vs multi-object) and compile_preflight vs await_compile (pre-write validation vs post-write wait). No two tools appear to overlap.
The dominant pattern is verb_noun snake_case (get_hierarchy, create_object, run_tests), with a few noun-only tools (scene, editor, batch) that function as grouped action commands. The inconsistency is minor but prevents a perfect score.
With 36 tools, this exceeds the 25-tool threshold for 'too many' per the rubric. While the domain is complex, the server could reduce surface area by consolidating related operations (e.g., console_mark+get_console_since, scene_change_plan+apply_scene_change).
The set covers scene hierarchy creation/deletion/editing, component inspection/modification, play mode testing, test dispatch, compilation, console logging, screenshots, and editor state. Minor gaps exist (e.g., no duplicate object, no asset creation beyond prefabs) but execute_code provides a general escape hatch.