Scored across 165 tools
Many tools have unclear boundaries, especially financial_report vs financial_statement (nearly identical IS/BS/CF interfaces), candlesticks vs history_candlesticks_by_date vs history_candlesticks_by_offset, and the cluster of institution_rating/shareholder tools. The broad domain coverage helps, but agents would frequently misselect between near-duplicate market-data and financial-data tools.
Names are consistently snake_case and many use readable domain prefixes like alert_, dca_, grid_, and sharelist_. However, the pattern is mixed: retrieval tools are bare nouns (quote, depth, dividend, company) while others use verb_noun (submit_order, cancel_order, create_watchlist_group), and oddities like now and operating break the convention.
165 tools is extreme bloat for a single server, far beyond the well-scoped 3-15 range. Even accounting for the many domains (trading, market data, IPOs, screener, community, macro), the tool surface is overwhelming and could be consolidated significantly.
The server covers an unusually wide breadth of brokerage and market-data workflows: order lifecycle, account operations, positions, deposits/withdrawals, alerts, DCA, grid trading, watchlists, sharelists, options, warrants, IPOs, screener, news, fundamentals, and macro data. There are minor dead ends such as no create/update/delete for screener user strategies and no IPO order submission/cancellation, but agents can generally work around these.