Stock and Fund Analysis Assistant

Problem

Investment research often involves fragmented data: fund NAV and performance are spread across pages, manual return/drawdown/rank calculations are error-prone, and stock quotes, candlesticks, sectors, and flow data live in separate sources. This skill turns public-market fund and stock analysis into repeatable report workflows for engineers.

How It Works

• Fund analysis: supports single-fund deep analysis, multi-fund comparison, portfolio return calculation, and backtesting, with stage returns, peer rankings, quartile ratings, risk checks, similar-fund recommendations, and market context.
• Stock analysis: supports minute candlesticks, sector/index quotes, batch stock comparison, order-book depth, large-order flow, auction anomalies, intraday trends, and natural-language stock screening.
• Data and fallback: fund data targets Tiantian Fund, with fund_api.py handling code resolution, NAV retrieval, historical NAV, fund details, holdings analysis, return and backtest calculation; stock data uses stock_api.py over thsdk for symbol resolution, real-time quotes, candlesticks, intraday data, and sector constituents.
• Report output: each mode should produce a self-contained HTML report using the A-share convention of red for gains and green for declines, with consistent styling for returns, rankings, and ratings.

Boundaries

It provides public-market data analysis only: no real-time trading, no personal account access, no future-return prediction, and no private-fund coverage. Confirmed NAV and intraday estimated NAV must be distinguished, non-trading days should use the prior trading day, and large batch analyses may time out.