Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股市场状态/牛熊识别量化。当用户说"市场状态"、"regime"、"牛市还是熊市"、"现在是什么市"、"震荡市"、"趋势市"、"牛熊判断"时触发。基于 cn-stock-data 获取K线数据，量化识别当前市场状态。支持研报风格（formal）和快速分析风格（brief）。
.claude/skills/aifinlab-a-share-regime-detection/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -39% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" python "$SCRIPTS/cn_stock_data.py" kline --code [CODE] --freq daily --start [日期] python "$SCRIPTS/cn_stock_data.py" quote --code [CODE] python "$SCRIPTS/cn_stock_data.py" finance --code [CODE]
获取主要指数（上证/沪深300/创业板）K线 + 成交量 + 资金流。
标注过去 N 年的市场状态序列，统计各状态持续时间。
| 维度 | formal | brief | |------|--------|-------| | 当前状态 | 多维度评分 | 一句话判断 | | 信号详情 | 各指标明细 | 关键信号 | | 历史对比 | 类似状态历史 | 无 |
默认风格：brief。
python# 调用 skill result = run_skill({ "param1": "value1", "param2": "value2" })
bashpython scripts/run_skill.py --input data.json
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | pass→fail | 26,460 | 10,247 | -61% | 1 | 1 | 0% | 5,266 | 1,229 | -77% | 0 | 0 | — |
case-21 | pass→pass | 31,536 | 32,671 | +4% | 1 | 1 | 0% | 5,075 | 5,797 | +14% | 0 | 0 | — |
case-01 | fail→fail | 9,907 | 10,062 | +2% | 1 | 1 | 0% | 1,335 | 1,036 | -22% | 0 | 0 | — |
case-02 | fail→fail | 31,499 | 114,701 | +264% | 1 | 1 | 0% | 4,886 | 3,897 | -20% | 0 | 0 | — |
case-03 | fail→fail | 29,212 | 10,807 | -63% | 1 | 1 | 0% | 4,799 | 1,393 | -71% | 0 | 0 | — |
case-04 | pass→pass | 25,897 | 22,740 | -12% | 1 | 1 | 0% | 4,653 | 3,813 | -18% | 0 | 0 | — |
case-05 | fail→fail | 24,852 | 24,268 | -2% | 1 | 1 | 0% | 3,106 | 3,623 | +17% | 0 | 0 | — |
case-06 | fail→pass | 12,749 | 5,231 | -59% | 1 | 1 | 0% | 2,341 | 1,443 | -38% | 0 | 0 | — |
case-07 | fail→pass | 13,285 | 4,539 | -66% | 1 | 1 | 0% | 2,359 | 1,406 | -40% | 0 | 0 | — |
case-08 | pass→pass | 16,464 | 5,061 | -69% | 1 | 1 | 0% | 2,083 | 1,145 | -45% | 0 | 0 | — |
case-09 | pass→pass | 24,340 | 21,800 | -10% | 1 | 1 | 0% | 3,121 | 3,558 | +14% | 0 | 0 | — |
case-10 | pass→pass | 18,201 | 19,932 | +10% | 1 | 1 | 0% | 2,308 | 2,902 | +26% | 0 | 0 | — |
case-11 | fail→pass | 20,265 | 9,157 | -55% | 1 | 1 | 0% | 2,841 | 1,911 | -33% | 0 | 0 | — |
case-12 | fail→pass | 16,844 | 5,507 | -67% | 1 | 1 | 0% | 2,155 | 1,299 | -40% | 0 | 0 | — |
case-13 | fail→pass | 11,585 | 2,464 | -79% | 1 | 1 | 0% | 1,559 | 955 | -39% | 0 | 0 | — |
case-14 | fail→pass | 22,119 | 22,872 | +3% | 1 | 1 | 0% | 2,658 | 4,069 | +53% | 0 | 0 | — |
case-15 | pass→pass | 19,823 | 8,954 | -55% | 1 | 1 | 0% | 2,427 | 1,674 | -31% | 0 | 0 | — |
case-16 | fail→pass | 19,368 | 17,189 | -11% | 1 | 1 | 0% | 3,239 | 3,394 | +5% | 0 | 0 | — |
case-17 | pass→pass | 14,615 | 5,763 | -61% | 1 | 1 | 0% | 2,118 | 1,548 | -27% | 0 | 0 | — |
case-18 | pass→pass | 11,910 | 3,133 | -74% | 1 | 1 | 0% | 1,897 | 1,044 | -45% | 0 | 0 | — |
case-19 | pass→pass | 24,742 | 15,335 | -38% | 1 | 1 | 0% | 3,788 | 3,207 | -15% | 0 | 0 | — |
case-22 | pass→pass | 35,959 | 30,730 | -15% | 1 | 1 | 0% | 4,502 | 6,539 | +45% | 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 19 counted toward the lift figure. The other 3 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 +27 percentage points is the difference between those two pass rates over the 19 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.