Industry Research Summary

Problem

Broker research reports are scattered across institutions, dates, and analytical angles. Manually reviewing them often means checking date ranges, prioritizing brokerages, and splitting original viewpoints into comparable categories. This skill turns that process into a query-to-structured-output flow, useful for building a quick industry-view snapshot without replacing investment judgment.

How It Works

The skill relies on mcporter and the jy-financedata-tool MCP service, with FinancialResearchReport as the core tool. It extracts beginDate, endDate, and industry keywords from the request; if no date range is specified, it defaults to the past week to keep results timely.

Queries prioritize the configured broker set first and expand to additional brokerages when needed. Priority matching happens in the background and is not exposed in the final output. Returned report fragments are deduplicated by doc_id or by title, brokerage, and publish date, then classified by keywords into Industry Trends, Competitive Landscape, Technology Innovation, Policy Environment, Investment Recommendations, Risk Warnings, Financial Data, and other categories. Results are sorted by publishDate in descending order and returned as JSON.

Boundaries

The skill preserves original viewpoints and does not freely rewrite or summarize them. Dimension classification depends on keyword rules, so unmatched content is grouped under Other. If no reports are found, the user may need to widen the date range, adjust industry keywords, or relax brokerage restrictions. Multi-industry comparisons usually require separate calls per industry rather than one complete comparative conclusion.