Yesterdays Number

Most domain clusters are distinct, and the descriptions explicitly route between similar tools such as entity_profile vs company_facts and ask_pipeworx vs deep_research. However, ask_pipeworx_beta is currently identical to ask_pipeworx, polymarket_arbitrage and polymarket_edges both surface partition/overround opportunities, and the onboarding tools overlap enough that selection requires careful reading.

Every name is lowercase snake_case, so there is no camelCase chaos, but the convention mixes bare verbs (remember, subscribe), verb_noun forms (resolve_entity, validate_claim), and noun phrases (entity_profile, polymarket_edges, company_facts). It remains readable and largely domain-prefixed, but it is not a single predictable pattern.

45 tools is well above the 3–15 well-scoped range and beyond the 25+ heavy threshold. The platform is broad, but several tools are experimental, duplicative, or narrow variants that could be consolidated, and the count creates a real selection burden for an agent.

For an all-in-one research/data gateway, coverage is extensive: memory and subscription lifecycle tools, SEC/company/entity data, web archives, AI visibility, npm dependency scanning, and deep prediction-market analytics. Minor gaps remain around generic live URL fetching/search and a few descriptions reference market or financial tools not exposed here, but agents can route around most of these through ask_pipeworx or deep_research.