Jungle Scout

The verbose descriptions do a good job of explaining when to use each tool, but several clusters blur together: the 6 polymarket_* tools (edges, arbitrage, fill_risk, kalshi_spread, edge_tracker, bet_research) all orbit the same "find an edge" task, and ask_pipeworx_beta currently behaves identically to ask_pipeworx. An agent could easily pick the wrong one despite the guidance.

All names are snak_case and readable, and cluster prefixes (polymarket_, resolution_, ask_pipeworx) help, but the verb/noun convention is mixed: compare_entities and discover_tools sit alongside entity_profile and company_facts, plus bare verbs like remember/forget/subscribe. There's a clear logic per cluster, but no single pattern across the set.

At 39 tools this exceeds the 'too many' threshold, and many could be consolidated: 3 ask_pipeworx variants, 6 Polymarket analytics tools, and 3 memory tools add surface area without much marginal capability. The scope is genuinely broad (universal data router + prediction markets + Amazon seller tools + subscriptions + meta-tools), but the count is bloated and the Jungle Scout name implies a much narrower purpose than the set delivers.

For the de-facto domain (Pipeworx-style structured data lookups + prediction-market analysis), coverage is strong: lookup, grounded answers, deep research, profiles, comparisons, resolution-audit, claim validation, subscriptions, and memory all form working lifecycles. But the server is named Jungle Scout and only includes 3 Amazon tools — no product tracker, opportunity finder, or review analysis — and unrelated utilities (scan_depency, generate_llms_txt, ai_visibility_check) feel tacked on.