Tastedive

Several tools occupy the same conceptual space: ask_pipeworx, ask_pipeworx_beta, ask_pipeworx_grounded, deep_research, discover_tools, and suggest_questions all serve overlapping query/discovery roles, and ask_pipeworx_beta is currently an exact duplicate of ask_pipeworx. The six polymarket_* tools and the two AI-visibility tools also blur boundaries, making misselection likely without deep description reading.

Almost all tool names are descriptive snake_case and many follow recognizable family prefixes like ask_pipeworx, polymarket_, pipeworx_, and the remember/recall/forget trio. Minor inconsistency exists between verb-led names (resolve_entity, validate_claim), noun-led names (entity_profile, pipeworx_trending), and the odd one-off get_recommendations, but there is no chaotic mixing of conventions.

With 32 tools, this exceeds the 25+ threshold for 'too many,' and the scope is sprawling across structured data query, prediction markets, memory, subscription management, AI visibility, npm dependency scanning, and TasteDive recommendations. A server named Tastedive carrying only one TasteDive-related tool suggests the set is heavily over-stuffed and not well-scoped.

Within its inferred broad purpose as a structured-data and prediction-market platform, the tool surface is impressively complete: query routing, grounded answers, deep research, entity resolution, comparisons, claim verification, change feeds, discovery, subscriptions, and memory are all represented. Minor gaps like a direct citation-URI fetch tool and the absence of source-specific direct-access tools keep it from a perfect score.