name: daily-a-share-digest
description: Use when preparing A股短线王 / daily A-share market research for mainland China equities: indexes, breadth, volume, hot themes, sector rotation, 连板天梯, 龙虎榜, 资金流向, major announcements, risk notes, optional watchlists, 风扇理论板块轮动, 蜂巢理论资金迁徙, repair/bottom-fishing scans, backtest-oriented observation, and position checks. Triggers for A股短线王, 每日A股动态合集, A股动态, A股复盘, 国内股票, 沪深京股票, 股票观察清单, 板块轮动, 修复低吸, 抄底观察, 风扇理论, 蜂巢理论, or next-session A股 tracking. This skill must not present personalized investment advice or guaranteed predictions.
A股短线王
Use this skill to produce A股短线王 research notes: market temperature, hot themes, sentiment, capital flow, announcements, sector-rotation models, optional watchlists, and risk-aware short-term scenarios. The output is a research aid, not personalized financial advice.
Safety Rules
- Do not promise returns, use deterministic language like "必涨", or claim to predict exact prices.
- Do not tell the user to buy, sell, hold, add leverage, all-in, or trade at any exact price as a directive.
- Label candidates as "观察标的" or "候选池", not "must-buy recommendations". A digest can omit stock candidates entirely when market conditions are poor.
- Include risk notes, invalidation conditions, and why the idea may fail.
- If the user asks for personal allocation, portfolio construction, suitability advice, or account-specific trade sizing, explain that this requires a licensed adviser and provide only general education.
- For current market data, news, rules, or prices, browse or use live data tools before answering.
- Treat rumors, social-media posts, and forum claims as unverified unless confirmed by company announcements, exchange disclosures, official filings, or reputable media.
User Constraints
Ask for or infer constraints before narrowing the watchlist:
- Broker/app access, such as 同花顺, 东方财富, or a specific securities account.
- Approximate cash scale and whether the user wants one-lot affordability filters.
- Confirmed permissions, such as 创业板, 科创板, 北交所, 港股通, options, margin trading, or fund products.
- Current positions and cost basis if the user wants position checks.
- Risk preference: conservative, balanced, aggressive, or short-term momentum.
If the user gives no constraints, default to ordinary Shanghai/Shenzhen main-board A-shares and avoid special-permission markets.
Retail Account And Budget Filter
When the user mentions a small cash budget or ordinary retail account, apply this default filter unless the user says otherwise:
- Assume the user can place ordinary A-share orders through their broker app, but do not assume special permissions.
- Prefer Shanghai/Shenzhen main-board A-shares that a normal retail account can usually access.
- Exclude 科创板, 北交所, margin/short candidates, B-shares, 港股通, ETFs, convertible bonds, and other products unless explicitly requested.
- Exclude 创业板 if the user has not confirmed they have 创业板 permission.
- Exclude ST/*ST, delisting-risk names, suspended names, names under major investigation, and obvious liquidity traps.
- A-share buy orders are commonly in 100-share lots. For small accounts, show "一手约需资金 = 最新价 x 100" for every candidate.
- Mark candidates that would consume most of the user's cash as "资金集中风险".
- If fewer than 3 quality candidates pass the filter, say so and provide fewer names rather than lowering risk controls.
Market Sentiment Add-On
When available, incorporate A-share short-term sentiment data before ranking candidates:
- Earning effect: market breadth, limit-up/limit-down count, fried-board rate, advancing vs declining names.
- Theme heat: top concepts/industries by涨幅, turnover, and persistence from morning to afternoon.
- 连板天梯 and high-board names: use them to judge whether risk appetite is expanding or contracting.
- 龙虎榜 and institutional/游资 clues: only use as supporting evidence, not a buy signal by itself.
- Capital flow: compare individual stock volume/turnover with sector heat; flag放量滞涨 and冲高回落.
- Trading calendar: distinguish trading days, holidays, pre-market, intraday, lunch break, and post-close.
The GitHub project Niceck/hhxg-top-hhxg-python can be used as a reference for market-sentiment categories such as赚钱效应、热门题材、连板天梯、龙虎榜、行业资金、财经快讯、融资融券、交易日历. Treat third-party data as unverified until cross-checked with Eastmoney, Tonghuashun, Sina Finance, exchange announcements, or company filings.
Rotation Models
Use these models as structured observation frameworks, not prediction engines. For details, load references/fan-honeycomb-models.md when the user asks about 风扇理论, 蜂巢理论, 板块周期, 修复低吸, or capital migration.
- 风扇理论: Treat sectors as fan blades. Classify each blade as 冷却、潜伏、启动、加速、高潮、分歧、退潮、修复 by price structure, turnover, breadth, limit-up feedback, leader/follower behavior, and invalidation lines.
- 蜂巢理论: Treat sectors as honeycomb cells and main funds as bees. Use half-year style evidence such as 5d/20d/60d/6m returns, range position, volatility, and volume to judge which cells are crowded, cooling, or attracting test flows.
- Repair/bottom-fishing rule: A candidate is not valid just because it fell. Prefer names that stop falling, reclaim VWAP/均价 or a short moving average, keep sufficient liquidity, and have a clear failure line.
- Scenario language only: Express conclusions as 强/中/弱情景 with confirmation and invalidation signals.
Bundled Tools
Use bundled scripts when local execution is helpful:
scripts/a_share_watch.py: fetch public quotes and render current price, one-lot cash, position P/L, key lines, and risk state.
- Example:
python scripts/a_share_watch.py -p 600000:100:10.00 000001
scripts/a_share_scan.py: scan ordinary Shanghai/Shenzhen main-board names from Eastmoney-style public data and rank repair/low-absorption candidates with transparent factors.
- Example:
python scripts/a_share_scan.py --pages 8 --top 30
apps/a_share_visual_app.py: Streamlit dashboard for Mac/browser use, including holdings, fan-blade rotation, honeycomb flow, and repair scan.
- Example:
streamlit run apps/a_share_visual_app.py
Daily Workflow
- Confirm date, market status, and whether A股 is open. Note if the brief is pre-market, intraday, lunch break, post-close, weekend, or holiday.
- Gather current sources:
- Official exchanges and regulators: SSE, SZSE, BSE, CSRC.
- Company filings and announcements: 巨潮资讯, exchange announcements, company investor relations pages.
- Market data portals: Eastmoney, Tonghuashun, Sina Finance, Tencent Finance, Wind-like summaries when accessible.
- News and catalysts: 证券时报, 财联社, 上海证券报, 中国证券报, official policy releases.
- Optional quantitative data: AkShare or Tushare if available locally; verify freshness before trusting it.
- Build the daily A股 dynamic digest:
- Major index trend: 上证指数, 深证成指, 创业板指, 科创50, 北证50.
- Breadth: 涨跌家数, 涨停/跌停, 量能 vs recent average.
- Northbound/ETF/fund flow when available.
- Sector rotation and policy/event catalysts.
- 热门题材: rank strongest themes and whether they are first-day, continuation, or late-stage hype.
- 连板天梯: summarize high-board height and whether risk appetite is expanding or contracting.
- 龙虎榜/机构/游资 clues: include only if timely and source-backed.
- 重要公告/财经快讯: list catalysts that matter for the next session.
- 风险温度: identify overheated sectors, rumor-heavy stories, and obvious liquidity traps.
- 风扇叶片: summarize leading sectors by rotation stage and next confirmation point when sector-cycle analysis is requested.
- 蜂巢迁徙: summarize which sectors look crowded, cooling, repairing, or attracting trial capital when medium-term analysis is requested.
- Check current positions when the user provides them:
- Current price vs cost basis.
- Key support/resistance and invalidation levels.
- Strong/base/weak next-session scenarios.
- Whether new candidates duplicate existing portfolio exposure.
- Optionally screen watchlist candidates with multiple, independent signals:
- Liquidity: avoid thinly traded names unless explicitly requested.
- Momentum: recent relative strength vs index and sector.
- Volume/price confirmation: breakout, reversal, or consolidation with clear invalidation.
- Fundamentals: profitability, revenue trend, valuation, balance-sheet stress, pledge or delisting risk.
- Catalysts: policy, earnings, order wins, industry events, announcements.
- Risk filters: ST/*ST, regulatory investigation, abnormal volatility, imminent lockup expiry, large shareholder reduction, repeated failed breakouts.
- Accessibility and budget: apply the user's retail account and one-lot affordability filter when relevant.
- Rank with a transparent score, not a black box:
- Trend and relative strength: 25%
- Volume and capital attention: 20%
- Catalyst clarity: 20%
- Fundamental/valuation support: 20%
- Risk control quality: 15%
- Forecast only as scenarios:
- Bull/base/bear paths for the next session or next 3-5 trading days.
- Assign rough probabilities only if evidence supports them.
- State invalidation signals such as sector breakdown, index volume failure, announcement risk, or key support loss.
Output Format
Write in Chinese unless the user asks otherwise.
Start with this disclaimer:
"以下是A股短线王研究观察，不构成个性化投资建议。"
Then provide:
- 市场温度: date, status, broad index read, breadth, volume, limit-up/limit-down count, and risk appetite.
- 热门题材: 3-6 themes or industries with rotation status and why they matter.
- 风扇叶片: optional sector-rotation table with stage, representative stocks, capital temperature, and next confirmation point.
- 蜂巢迁徙: optional medium-term capital migration read, especially when the user asks for half-year patterns or next sector rotation.
- 连板与情绪: high-board height, fried-board clues, 龙虎榜/资金 clues when available.
- 公告与快讯: major policy, company, and macro catalysts relevant to A股.
- 持仓检查: only if the user provided positions; include current price, cost line, support, failure line, and next-session scenarios.
- 观察清单: optional, 0-8 stocks max, each with:
- 股票名称 + 代码
- 所属行业/概念
- 最新价, 一手约需资金, and whether it fits the user's stated cash budget
- 关注理由, using current evidence
- 情景推演: 强/中/弱 path
- 风险与失效条件
- 观察级别: 高/中/低, based on evidence quality
- 不碰清单: obvious risks or overheated areas.
- 下个交易日跟踪点: 3-5 concrete signals to verify.
Keep it concise, source-backed, and timestamped. Link the most important sources when browsing was used.
Recommended Sources To Check
