Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股行业轮动策略/风格切换分析。当用户说"行业轮动"、"轮动策略"、"风格切换"、"现在轮到什么行业"、"sector rotation"、"大小盘切换"、"成长价值切换"、"什么行业有机会"、"下一个风口"时触发。基于 cn-stock-data 行情数据和宏观经济周期，分析当前行业轮动阶段和风格特征，提示潜在切换方向。支持策略报告风格（formal）和快速判断风格（brief）。
.claude/skills/aifinlab-a-share-sector-rotation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | 9% | 0% |
| case-08 | ✓→✓ | ＝ Same ✓ | 48% | 0% |
| case-09 | ✓→✓ | ＝ Same ✓ | 41% | 0% |
使用 cn-stock-data skill 获取以下数据：
# 行业轮动策略报告
## 经济周期定位
[当前阶段判断及依据]
## 风格特征
| 维度 | 当前状态 | 趋势方向 | 信号强度 |
|------|---------|---------|---------|
| 大盘/小盘 | ... | ... | ... |
| 价值/成长 | ... | ... | ... |
## 行业强弱格局
[近5/10/20日排名表格，标注动量变化]
## 轮动信号与配置建议
- 看好行业：[行业] — [逻辑]
- 警惕行业：[行业] — [逻辑]
- 潜在切换：[从X到Y的可能路径]
## 风险提示
[政策/外部事件/数据验证风险]周期阶段：[X]，风格偏[大/小盘]+[价值/成长]
强势行业：A > B > C（近10日动量）
弱势行业：X < Y < Z
轮动信号：[一句话核心判断]
建议：[2-3句操作建议]| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 27,293 | 9,983 | -63% | 1 | 1 | 0% | 3,862 | 1,505 | -61% | 0 | 0 | — |
case-02 | fail→fail | 24,809 | 37,194 | +50% | 1 | 1 | 0% | 2,008 | 1,516 | -25% | 0 | 0 | — |
case-03 | fail→fail | 28,103 | 15,487 | -45% | 1 | 1 | 0% | 3,951 | 1,978 | -50% | 0 | 0 | — |
case-04 | fail→fail | 30,315 | 43,113 | +42% | 1 | 1 | 0% | 5,290 | 7,273 | +37% | 0 | 0 | — |
case-05 | fail→fail | 37,437 | 11,757 | -69% | 1 | 1 | 0% | 5,371 | 1,688 | -69% | 0 | 0 | — |
case-06 | fail→fail | 28,308 | 28,201 | -0% | 1 | 1 | 0% | 4,133 | 4,877 | +18% | 0 | 0 | — |
case-07 | pass→pass | 18,654 | 15,094 | -19% | 1 | 1 | 0% | 3,207 | 3,510 | +9% | 0 | 0 | — |
case-08 | pass→pass | 31,719 | 34,905 | +10% | 1 | 1 | 0% | 3,263 | 4,829 | +48% | 0 | 0 | — |
case-09 | pass→pass | 11,551 | 11,840 | +3% | 1 | 1 | 0% | 1,785 | 2,509 | +41% | 0 | 0 | — |
case-10 | fail→pass | 25,760 | 18,099 | -30% | 1 | 1 | 0% | 3,128 | 3,515 | +12% | 0 | 0 | — |
case-11 | fail→pass | 20,580 | 27,434 | +33% | 1 | 1 | 0% | 2,806 | 4,259 | +52% | 0 | 0 | — |
case-12 | pass→pass | 24,402 | 35,039 | +44% | 1 | 1 | 0% | 3,121 | 5,037 | +61% | 0 | 0 | — |
case-13 | pass→pass | 20,021 | 26,100 | +30% | 1 | 1 | 0% | 2,747 | 3,955 | +44% | 0 | 0 | — |
case-14 | fail→fail | 18,042 | 5,986 | -67% | 1 | 1 | 0% | 2,774 | 1,941 | -30% | 0 | 0 | — |
case-15 | fail→fail | 12,123 | 21,787 | +80% | 1 | 1 | 0% | 1,893 | 1,613 | -15% | 0 | 0 | — |
case-16 | pass→pass | 17,666 | 9,104 | -48% | 1 | 1 | 0% | 2,307 | 2,277 | -1% | 0 | 0 | — |
case-17 | fail→fail | 22,767 | 23,202 | +2% | 1 | 1 | 0% | 2,879 | 4,004 | +39% | 0 | 0 | — |
case-18 | fail→fail | 29,725 | 14,807 | -50% | 1 | 1 | 0% | 4,879 | 1,647 | -66% | 0 | 0 | — |
case-19 | pass→pass | 30,029 | 15,350 | -49% | 1 | 1 | 0% | 2,541 | 2,868 | +13% | 0 | 0 | — |
case-20 | fail→fail | 23,211 | 21,151 | -9% | 1 | 1 | 0% | 2,987 | 3,677 | +23% | 0 | 0 | — |
case-21 | pass→pass | 21,371 | 6,344 | -70% | 1 | 1 | 0% | 3,492 | 1,864 | -47% | 0 | 0 | — |
case-22 | pass→pass | 18,148 | 2,719 | -85% | 1 | 1 | 0% | 2,397 | 1,401 | -42% | 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 16 counted toward the lift figure. The other 6 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 +9 percentage points is the difference between those two pass rates over the 16 comparable cases. 1 case got worse with the skill loaded, and it is 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.