Hundsun Gildata Financial Data

What problem it solves

Finance queries often mix precise numbers, interpretive views, primary-source verification, and candidate screening in one request. For example, asking about a stock may require quote data, financials, valuation, announcements, research notes, and sentiment. If the agent simply assembles data, it can introduce inconsistent definitions, replace structured values with research narratives, or confuse single-asset lookup with bulk screening.

This skill routes finance-data questions to Gildata MCP tools by data type: FinQuery for structured data on known instruments, MacroIndustryData for macro and industry time series, AnnouncementData for company announcements, FinancialResearchReport and InvestMeetingNoteQuery for research views and investment meeting notes, SmartStockSelection, SmartFundSelection, and SmartFundManagerSelection for candidate screening, NewsDataQuery for news and sentiment, IcEnterpriseDataQuery for enterprise and compliance data, and FinDataFallbackQuery as a fallback interface.

Key workflow and boundaries

• Route before calling: identify whether the user needs a single-asset detail, a candidate list, a time series, announcement text, or narrative analysis, instead of trial-and-error tool selection.
• Normalize assumptions: use defaults only when ambiguity is low, and state the assumptions clearly; when ambiguity is high, clarify the instrument, time range, and metric definition first.
• Prefer structured data: precise values should come from structured tools, not research reports or news snippets.
• Use minimal tools: call the fewest tools that can preserve answer quality, reducing mismatched definitions.
• Keep boundaries clear: index constituents use screening tools; index quotes or valuation use FinQuery; company disclosures use announcement data; analyst views use research or meeting notes; enterprise, litigation, guarantee, and credit-risk records use enterprise data.