4 eli-alvarado

watchlist-brief Skill

生成美股自选清单的盘前(早间)或收盘后简报——价格与涨跌幅、RSI、趋势、距止损和目标价的距离、涨跌幅榜、即将到来的财报以及触发提醒。当用户要求"早报"、"行情回顾"、"今天自选表现如何",或在定时早间/收盘任务时使用。

安装方式:把技能目录放入 ~/.claude/skills/(Claude Code)或在 claude.ai 设置中启用;也可复制右侧安装命令一键添加。

查看源码

技能指令原文(SKILL.md)

Watchlist Brief (morning / close)

Informational summary only - not financial advice; no trade suggestions or order placement.

Steps

scripts/ ships next to this SKILL.md after npx skills add. Run them from the project root by path (do not cd into this skill folder). Same shared watchlist as the other two skills: $WATCHLIST_FILE, else ./watchlist.json in the project directory (walks up from .claude/skills/ / .agents/skills/), else ~/.config/us-watchlist-alerts/watchlist.json.

  1. Determine context with python3 scripts/market_calendar.py. Morning brief ~08:30-09:25 ET; close brief after 16:05 ET. On holidays/weekends, say the market is closed and give the next session. After the bell the alerter evaluates last regular close and labels it — do the same in the brief.
  2. Generate the table: python3 scripts/brief.py --kind morning (or --kind close). Columns: Ticker | Price | Chg | RSI14 | Trend (SMA50 vs SMA200) | To stop | To target 1. "To stop" is remaining cushion (positive = room left); shorts use stop-minus-price. Earnings dates in the footer are upcoming only (Yahoo calendar/HTML, Nasdaq analyst page, then Finnhub/AV if keyed). Use --format json when a downstream script will consume the brief. Biggest movers only: python3 scripts/movers.py. Gaps vs prior close: python3 scripts/gappers.py --threshold 2.
  3. Run python3 scripts/alerts.py --dry-run --format json to list conditions currently true without consuming cooldowns; mention notable ones. After hours this uses last regular close (no --force-session). Use --force-session only on weekends/holidays if you still want to evaluate.
  4. Optionally chart a notable ticker: python3 scripts/chart.py AAPL --watchlist watchlist.json (or --style candle --sma 20,50). Stdlib ASCII sparkline + grid with SMA overlays and stop / entry / target / price_above / price_below lines from the watchlist. --out ticker.png writes a PNG when matplotlib is installed; otherwise it writes a .txt next to that path. Informational only — never a trade signal.
  5. Optionally add 1 headline per ticker that moved > 3% or has earnings within 7 days (web search, or tickertick/news tools if available). Cite sources; never fabricate news.
  6. Write the brief:
  • Headline (one sentence: the overall picture of the list).
  • The table.
  • Watch today / tomorrow: earnings, tickers within 3% of stop or target, gap moves.
  • Data source + delay note, and "Informational only - not financial advice."

Keep it under ~250 words unless asked for more.

Morning vs close

  • Morning: use prior close as baseline; mention pre-market moves only if the data source returns them (label as pre-market).
  • Close: day's change, closing levels vs stops/targets, anything that triggered today, what's on tomorrow's calendar.

Tone

Neutral and factual: "TSLA is 2.1% above its stop" - not "you should sell TSLA".