Tides

Tools cluster into overlapping groups (data lookup, prediction markets, memory, subscriptions), with several near-duplicates: ask_pipeworx and ask_pipeworx_beta are currently identical, and ask_pipeworx_grounded vs validate_claim both verify factual claims with evidence. Detailed descriptions and explicit 'use this instead' guidance mitigate some confusion, but an agent could still struggle to choose between the many research/entity tools.

Naming is a mixed bag: verb_first (compare_entities, generate_llms_txt), noun_first (ai_visibility_check, entity_profile), single verbs (remember, forget, subscribe), and branded prefixes (ask_pipeworx*, polymarket_*). The polymarket_ prefix is consistent, but the overall set lacks a predictable verb_noun pattern across all tools.

At 31 tools, the server is on the heavy side, though the broad data-platform scope (structured lookups, research, prediction markets, memory, subscriptions) partially justifies the number. Several tools feel peripheral to the core mission (generate_llms_txt, scan_dependency, ai_visibility_check) and could be split into separate servers.

Core workflows are well covered: universal lookup (ask_pipeworx), grounded verification (ask_pipeworx_grounded, validate_claim), deep research, entity profiles, comparisons, prediction-market arbitrage/edges/fill-risk, memory, and subscriptions. Minor gaps exist (e.g., no direct tool for reading a pipeworx:// URI, and USPTO patents soft-fail), but the surface itself is broad and self-discoverable.