Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股市场情绪综合研判/恐贪指数/情绪面分析。当用户说"市场情绪"、"情绪面"、"恐慌"、"贪婪"、"sentiment"、"现在市场情绪怎么样"、"市场热度"、"赚钱效应"、"亏钱效应"、"市场温度"、"是不是过热了"、"恐贪指数"、"情绪综合研判"时触发。MUST USE when user asks about market sentiment, fear/greed index, market temperature, or whether the market is overheated/panicking. 聚合多维数据（涨跌比/换手率/涨停数/北向资金/融资余额/成交量等），构建市场情绪综合判断。支持研报风格（formal）和快速解读风格（brief）。
.claude/skills/aifinlab-a-share-sentiment/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -71% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -63% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 53% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | 16% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" # 大盘行情（涨跌数据、成交量） python "$SCRIPTS/cn_stock_data.py" quote --code SH000001,SZ399001,SZ399006 # 北向资金 python "$SCRIPTS/cn_stock_data.py" north_flow # 大盘K线（近期趋势，用于历史对比） python "$SCRIPTS/cn_stock_data.py" kline --code SH000001 --freq daily --start $(date -v-30d +%Y-%m-%d) # 资金流向（全市场） python "$SCRIPTS/cn_stock_data.py" fund_flow --code SH000001
补充数据（通过 akshare 或 web 搜索获取）：
从多个维度获取市场情绪数据，优先使用 cn-stock-data 脚本，不足部分用 akshare 或 web 搜索补充。
将各维度归一化为 0-100 分，加权平均得到综合评分：
详细评分体系见 references/sentiment-indicators.md。
当前情绪在近1年中的百分位位置，参考历史极端值校准。
| 维度 | formal（研报风格） | brief（快速解读） | |------|-------------------|------------------| | 输出 | 完整市场情绪报告 | 情绪温度计 + 核心结论 | | 指标 | 8-10个维度逐一分析 | 综合评分 + 3个关键指标 | | 历史对比 | 与近1年百分位对比 | 简单高/低判断 | | 结论 | 客观描述（"情绪处于偏乐观区间"） | 可加判断 |
默认使用 formal 风格；用户要求"简单说说"、"快速看下"时用 brief。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→fail | 34,903 | 10,201 | -71% | 1 | 1 | 0% | 5,334 | 1,540 | -71% | 0 | 0 | — |
case-05 | pass→fail | 23,811 | 16,151 | -32% | 1 | 1 | 0% | 4,568 | 1,669 | -63% | 0 | 0 | — |
case-01 | fail→fail | 26,622 | 12,265 | -54% | 1 | 1 | 0% | 3,848 | 1,505 | -61% | 0 | 0 | — |
case-02 | fail→fail | 14,710 | 13,708 | -7% | 1 | 1 | 0% | 2,067 | 1,747 | -15% | 0 | 0 | — |
case-03 | fail→fail | 20,689 | 12,679 | -39% | 1 | 1 | 0% | 3,090 | 1,435 | -54% | 0 | 0 | — |
case-06 | pass→pass | 22,715 | 34,583 | +52% | 1 | 1 | 0% | 3,471 | 5,314 | +53% | 0 | 0 | — |
case-07 | pass→pass | 20,232 | 19,566 | -3% | 1 | 1 | 0% | 2,892 | 3,367 | +16% | 0 | 0 | — |
case-08 | pass→pass | 15,733 | 9,227 | -41% | 1 | 1 | 0% | 2,122 | 2,306 | +9% | 0 | 0 | — |
case-09 | fail→pass | 13,802 | 10,242 | -26% | 1 | 1 | 0% | 1,883 | 2,341 | +24% | 0 | 0 | — |
case-10 | pass→pass | 30,035 | 17,395 | -42% | 1 | 1 | 0% | 1,923 | 2,493 | +30% | 0 | 0 | — |
case-11 | pass→pass | 10,354 | 8,001 | -23% | 1 | 1 | 0% | 1,590 | 2,074 | +30% | 0 | 0 | — |
case-12 | pass→pass | 15,846 | 19,986 | +26% | 1 | 1 | 0% | 2,154 | 3,475 | +61% | 0 | 0 | — |
case-13 | pass→pass | 18,179 | 26,268 | +44% | 1 | 1 | 0% | 2,551 | 3,593 | +41% | 0 | 0 | — |
case-14 | pass→pass | 27,426 | 13,107 | -52% | 1 | 1 | 0% | 3,262 | 2,813 | -14% | 0 | 0 | — |
case-15 | fail→fail | 19,804 | 11,183 | -44% | 1 | 1 | 0% | 2,725 | 2,332 | -14% | 0 | 0 | — |
case-16 | fail→fail | 18,329 | 11,857 | -35% | 1 | 1 | 0% | 2,649 | 1,306 | -51% | 0 | 0 | — |
case-17 | pass→pass | 16,298 | 29,606 | +82% | 1 | 1 | 0% | 2,294 | 3,332 | +45% | 0 | 0 | — |
case-18 | pass→pass | 18,676 | 18,725 | +0% | 1 | 1 | 0% | 2,733 | 3,510 | +28% | 0 | 0 | — |
case-19 | fail→fail | 15,559 | 9,629 | -38% | 1 | 1 | 0% | 1,873 | 1,421 | -24% | 0 | 0 | — |
case-20 | fail→fail | 30,458 | 9,015 | -70% | 1 | 1 | 0% | 3,529 | 1,318 | -63% | 0 | 0 | — |
case-21 | pass→pass | 6,461 | 6,541 | +1% | 1 | 1 | 0% | 1,192 | 2,029 | +70% | 0 | 0 | — |
case-22 | pass→pass | 25,431 | 9,021 | -65% | 1 | 1 | 0% | 2,169 | 1,963 | -9% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 14 counted toward the lift figure. The other 8 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of -5 percentage points is the difference between those two pass rates over the 14 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.