What problem does it solve? Investors relying on secondhand summaries (news, analyst notes) miss the context and signals hidden in original filings. This Skill reads primary sources—10-K/10-Q filings, earnings call transcripts, shareholder letters—and interprets what the numbers and management language actually mean for a single company's quarter or year. ## Core Features & Use Cases - Primary-source earnings analysis: Fetches filings from SEC EDGAR, HKEX, CNINFO, and IR pages, then extracts income statement, cash flow, and balance sheet changes with YoY comparisons. - Management tone and promise tracking: Flags candid vs. evasive language in MD&A and call Q&A, and checks whether prior guidance was delivered. - Footnote mining and anomaly detection: Checks related-party transactions, dilution, receivables/inventory growth vs. revenue, and cash-flow-to-earnings quality. - Data rigor and audit gate: Cross-validates key figures across sources with financial_rigor.py, generates SVG trend charts, and runs a 15% data sampling audit via report_audit.py before the report can be published. - Use Case: Ask for an earnings review of Tencent 2025Q4 and receive a structured report covering core financials, management tone, hidden footnote items, and a clear verdict on whether the quarter strengthens or weakens the investment thesis. ## Quick Start Ask the agent to run an earnings review on a company and period, for example: do an earnings deep-dive on Tencent 2025Q4 using original filings and the earnings call transcript.
