report_generation

You can automate weekly equity research report generation by providing a stock symbol and target week ending date. The system uses MCP tools to read offline data via DuckDB, assembling 16 structured metrics and context fields into a standardized Markdown document and summary JSON without external API calls.

An automated equity weekly report includes 16 mandated metrics such as alpha, momentum, and beta, alongside context fields. These metrics are sourced from offline data using get_weekly_metrics, list_news, list_filings, and list_peers to ensure consistent structure and data provenance for financial analysis.

The report generation workflow strictly enforces no external API calls, relying entirely on offline data via the report_generation_mcp server. It uses list_news, list_filings, and get_filing_section to gather context, ensuring all results are produced through upsert_report.py without ad-hoc data execution.

Yes, you can use DuckDB with pandas to generate financial analysis reports by reading offline data through MCP tools. This combination supports the extraction of 16 weekly metrics and peer context, building structured output via upsert_report.py for persistent storage and review.

To persist a generated weekly research report, all results must be produced via upsert_report.py. This script handles the storage of the complete weekly report, including the Markdown document and run-level summary JSON, ensuring the output is saved for subsequent review.

The limitation of using offline MCP tools for equity research is the strict enforcement of no external API calls, meaning all data must pre-exist locally. Additionally, no ad-hoc data or code should be executed outside the mandated MCP workflow using upsert_report.py.