Related Servers

Most tools have detailed, distinct purposes, but several pairs blur together: context_pack vs league_overview (both return league overview, strategy, proposals, freshness), and the memory/proposal cluster (journal_append, record_review, record_owner_decision, proposal_steps, confirm_done, create_proposal) requires careful reading to route correctly. The signal/trade-advice cluster (signals, buy_sell, find_trades, evaluate_trade, evaluate_add_drop, waiver_decision, streaming_plan) is heavily overlapping in intent. Descriptions mitigate much of this, so it lands mid-range rather than poor.

Overwhelmingly consistent snake_case throughout, with a clear verb_noun family (refresh_data, create_proposal, evaluate_trade, draft_pick, record_review) and a noun-only family for resource reads (my_roster, rankings, free_agents, standings, scoreboard). The only minor deviation is mixing noun-style and verb-style tools, but the style is predictable and readable.

54 tools is well past the comfortable range and several could clearly be consolidated (context_pack/league_overview, multiple proposal-status tools, signals/buy_sell/signal_grades). While the fantasy-GM domain is genuinely broad (draft, waivers, trades, calibration, journaling), the surface is bloated enough to strain tool selection. This is on the heavy end the rubric flags as 'too many'.

Coverage is strong across the lifecycle: data sync/health, roster/lineup, waivers, trades, drafts, proposals with review/approval/confirmation, journaling, and calibration/retro. The main deliberate gap is that Yahoo writes are manual by design (proposals return steps for the owner), which is a documented boundary rather than a defect. Minor gaps remain around anything requiring automated league writes.