fakehank/stock_qualitative_analysis_skills

Overview

This skill produces buy-side style, evidence-first qualitative stock analysis reports from company filings and public sources. It structures output chapter-by-chapter with explicit conclusions, detailed analysis, and strict citations, and inserts clear placeholders where evidence is missing. The default language is Chinese, but English output is available on request.

How this skill works

The agent ingests user-provided filings (PDF/HTML) and, when authorized, fetches SEC EDGAR documents. It extracts key sections (e.g., Item 1, 1A, 7, 8) and iterates through a template to generate each report section as Conclusion / Details / Evidence. Any factual claim is tied to a source string; missing facts receive an explicit placeholder describing required evidence.

When to use it

• You need an evidence-based qualitative analysis (buy-side style) rather than a numerical model.
• You want a structured report that maps directly to SEC filings and other cited public sources.
• You require strict non-hallucination guarantees and placeholders for missing evidence.
• You must produce bilingual output (Chinese default) or English on request.
• You need progressive, chapter-by-chapter outputs for collaborative review.

Best practices

• Provide company name and ticker/exchange and allow EDGAR fetch if you want remote filings included.
• Supply local PDFs or HTML filings first; the agent will prioritize local sources before remote fetching.
• Specify the time window (e.g., FY2018–FY2024) and language preference up front.
• Review and confirm the source manifest before full analysis if you need strict source control.
• Treat the report as qualitative evidence synthesis, not investment advice; verify data freshness when required.

Example use cases

• Create a Chinese-language qualitative diligence report for a US-listed company using the last three 10-K/10-Q filings.
• Generate a chaptered analysis highlighting product risks and competitive positioning with direct citations to MD&A and Risk Factors.
• Produce a compliance-ready source list and flagged placeholders where required evidence (e.g., segment disclosures) is missing.
• Prepare an English-version report for internal distribution that mirrors the Chinese template structure.
• Iteratively build a report across multiple sessions, saving each completed section to a report file.