No user-submitted related servers found.
Scored across 21 tools
Prepare/execute pairs are explicitly paired and each targets a distinct action (create, cancel, late cancel, save result, delete result), so boundaries are mostly clear. The riskiest overlaps are the read-side trio of get_upcoming_bookings / get_booking_history / get_personal_activity and the 1RM trio (get_exercise_1rm / find_exercise_1rm / get_exercise_rm_progression), which descriptions do distinguish but an agent could still misselect.
All names use snake_case with a predictable verb-first pattern (get_/find_/prepare_/execute_) plus a resource noun, with suffixes like _range for variants. Minor deviation is that prepare_* tools are actually read-only previews while get_* is used for other reads, so the verb doesn't perfectly signal read vs. stage.
21 tools is on the heavy side for a single gym-app integration, inflated by four prepare/execute safety pairs (bookings, cancellation, exercise results, deletion). The domain is broad enough (sessions, booking, workouts, 1RM, activity) that most tools earn their place, but it sits at the borderline of heavy.
Booking lifecycle, cancellation with late-cancellation credit handling, schedules, participants, published workouts, 1RM data, and personal activity are all covered with both preview and execution paths. Gaps are minor: no waitlist join/leave despite waitlisted states being reported, and no way to edit the booking plan or preferences (only read via get_account_context).