China Finance Analyzer

Problem

Financial chat answers can mix stale NAV estimates, out-of-sync portfolio copies, and hand-written HTML without charts. This skill turns China market analysis into a reproducible data pipeline for stocks, funds, and portfolio P&L.

How it works

• Data collection: scripts/analyze.collect(code) pulls Tencent quotes, historical K-lines, Eastmoney financials, and news; HK data uses Tencent endpoints.
• Agent analysis: the skill does not embed an LLM; the current conversation evaluates market context, company events, fundamentals, and technicals, then outputs a risk-controlled conclusion.
• Report rendering: for a single eligible stock, render_html() generates a light red/white/gray report with a K-line chart; sectors, active funds, and indexes skip HTML generation.
• Holdings and funds: daily_portfolio.py fetches fund estimate P&L, while screenshots or official NAVs correct stale or inaccurate estimates; gold/commodity funds prioritize date-checked official NAVs.

Boundaries

Best for A-share, HK-stock, fund valuation, portfolio P&L, and reanalysis workflows. It is not a substitute for sector-level conclusions. Sync the PORTFOLIO holdings list, treat fundgz estimates as potentially delayed, and keep A-share displays red for gains and green for declines.