No user-submitted related servers found.
Scored across 117 tools
Most tools have a distinct purpose and the descriptions aggressively cross-reference each other (e.g. "use gma3_pool_usage instead of gma3_objects 'Macro Thru'", fallbacks labelled as fallbacks), which rescues a lot. But there are four or five parallel read paths for the same show data (gma3_get_object, gma3_objects, gma3_list_children, gma3_dump, gma3_pool, plus typed readers like gma3_sequences/gma3_cues/gma3_macros) and several overlapping writer families (gma3_set_attribute/set_attributes/set_color/set_position, gma3_store_cue/store_cues/store_cue_part), so an agent must read carefully to avoid misselection.
Everything is snake_case under one gma3_ prefix, with a readable convention: noun names for read-only listings (gma3_sequences, gma3_fixtures, gma3_macros, gma3_tags) and verb_noun for mutations (gma3_store_preset, gma3_configure_executor, gma3_delete_tag). Minor deviations exist (gma3_get_object vs gma3_objects vs gma3_dump, gma3_preset singular for a read tool), but the pattern is predictable.
117 tools is far past any practical ceiling for an agent's tool selection, even granted that grandMA3 is a genuinely huge domain (patching, cues, phasers, macros, timecode, DMX output, OSC, layouts, tags). Several earn their place, but there is real duplication (dump vs get_object, command/lua fallbacks duplicating structured tools, multiple store_cue variants) that could be consolidated. The surface should be split into sub-servers or grouped namespaces.
Coverage is unusually deep: full lifecycle for cues, presets, macros, tags, timecode, layouts, patch, fixtures/positions, DMX/OSC configuration and show files, with read-back and rollback semantics. Gaps are minor but present: no dedicated delete/rename for sequences, groups or presets (only via the raw gma3_command fallback or gma3_set_property), and no layout or layout-element deletion.