Multiple tool families blur together: whale_activity, whale_daily_summary, and whale_movements_summary all return aggregated whale counts/totals with subtly different granularity, and the analyst family (analysts_signals, analysts_signals_all, analyst_archive, analyst_monthly_performance, signal_history_monthly) offers overlapping scopes that are hard to distinguish. The eight near-identical _monthly archive tools compound the problem. Even with detailed descriptions, an agent would struggle to reliably pick the correct tool.

The dominant pattern (resource + tier suffix like _history, _monthly, _summary) is applied fairly systematically, e.g. fear_greed/fear_greed_history/fear_greed_monthly. However, the analyst family inconsistently flips between analysts_ and analyst_ (analysts_signals vs analyst_archive vs analyst_top_tokens), analysts_signals_all uses an awkward _all suffix, and bridge_flows (live) vs bridge_flow_history (historical) pluralizes inconsistently.

56 tools is far too heavy for an MCP surface, even for a broad crypto-intelligence domain. The count is inflated by systematically triplicating every metric across live/history/monthly variants, plus redundant endpoints like analysts_signals_all that duplicate comma-separated multi-analyst support in analysts_signals. Many of these could be consolidated through query parameters.

Coverage is extensive across whales, analysts, sentiment, funding rates, stablecoins, gas, narratives, leverage, arbitrage, news, tax, and airdrops, with multi-tier temporal depth for most series. However, the tier pattern is inconsistent: arbitrage and whale scores lack live snapshots, there is no live market-pulse tool despite market_pulse_history, and the public whale movements archive omits explorer URLs, creating verification dead ends.