AgentSkillsCN

update

获取最新股票数据、运行估值、生成仪表盘，并启动实时服务器。当用户希望更新数据并查看仪表盘时使用。

SKILL.md
--- frontmatter
name: update
description: Fetch fresh stock data, run valuations, generate dashboard, and start the live server. Use when the user wants to update data and view the dashboard.
disable-model-invocation: true
argument-hint: "[--universe sp500] [--skip-fetch] [--skip-gbm] [--skip-nn] [--skip-classic] [--skip-scanner]"
allowed-tools: Bash

Update Data & Launch Dashboard

Run the full investment analysis pipeline and serve the dashboard.

Steps

  1. •

    Start the dashboard server in the background so the user can see progress immediately:

    bash
    uv run python neural_network/invest_training_package/scripts/dashboard_server.py &

    Tell the user: Dashboard is live at http://localhost:8080

  2. •

    Run the full update pipeline. Pass through any arguments the user provided (universe, skip flags):

    bash
    uv run python scripts/update_all.py $ARGUMENTS

    This runs (in order, unless skipped):

    • •Data fetching from yfinance
    • •GBM predictions (6 variants)
    • •Neural network multi-horizon predictions
    • •Classic valuations (DCF, RIM, ratios, etc.)
    • •Dashboard HTML regeneration
    • •Opportunity scanner

    The default universe is sp500. Common alternatives: international, japan, tech, growth, europe, spain, all, cached.

  3. •

    Report completion — summarize what ran, how long it took, and remind the user the dashboard is at http://localhost:8080

Important

  • •Always use uv run python for all commands
  • •Run the dashboard server FIRST so the user gets the link immediately
  • •The update pipeline can take 10-20 minutes for a full run; keep the user informed of progress
  • •If the dashboard server is already running on port 8080, skip starting it again and just run the update pipeline
  • •If the user only says /update with no arguments, default to --universe sp500