Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股跨资产配置/大类资产轮动策略。当用户说"跨资产"、"大类资产"、"资产配置"、"asset allocation"、"股债轮动"、"股债商"、"美林时钟"时触发。基于 cn-stock-data 获取数据，进行跨资产配置分析。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-cross-asset/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 13% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 14% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | 43% | 0% |
| case-08 | ✓→✓ | ＝ Same ✓ | 39% | 0% |
通过 cn-stock-data skill 获取数据：
# 跨资产配置报告
## 一、宏观周期
| 指标 | 数值 | 信号 |
|------|------|------|
## 二、资产表现
[各资产收益/波动/Sharpe]
## 三、配置建议
[各资产权重与调整方向]
## 四、风险评估## 跨资产速览
- 当前周期：复苏早期
- 建议：超配股票(50%)、标配债券(30%)、低配商品(10%)、现金(10%)
- 股债相关性 -0.2，分散化有效
- 关注：CPI回升可能触发周期切换参考 references/cross-asset-guide.md 获取详细方法论与 A股实证研究。
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-01 | fail→fail | 31,113 | 37,181 | +20% | 1 | 1 | 0% | 4,459 | 6,520 | +46% | 0 | 0 | — |
case-02 | fail→pass | 17,587 | 19,046 | +8% | 1 | 1 | 0% | 2,511 | 3,156 | +26% | 0 | 0 | — |
case-03 | fail→fail | 34,809 | 38,113 | +9% | 1 | 1 | 0% | 4,524 | 5,672 | +25% | 0 | 0 | — |
case-04 | pass→pass | 23,237 | 20,680 | -11% | 1 | 1 | 0% | 3,478 | 3,928 | +13% | 0 | 0 | — |
case-05 | fail→fail | 39,961 | 32,877 | -18% | 1 | 1 | 0% | 4,922 | 4,946 | +0% | 0 | 0 | — |
case-06 | pass→pass | 22,898 | 27,723 | +21% | 1 | 1 | 0% | 4,002 | 4,566 | +14% | 0 | 0 | — |
case-07 | pass→pass | 18,461 | 22,059 | +19% | 1 | 1 | 0% | 2,837 | 4,071 | +43% | 0 | 0 | — |
case-08 | pass→pass | 17,516 | 25,059 | +43% | 1 | 1 | 0% | 3,117 | 4,327 | +39% | 0 | 0 | — |
case-09 | pass→pass | 26,577 | 25,208 | -5% | 1 | 1 | 0% | 3,818 | 4,457 | +17% | 0 | 0 | — |
case-10 | pass→pass | 17,679 | 16,925 | -4% | 1 | 1 | 0% | 2,376 | 3,111 | +31% | 0 | 0 | — |
case-11 | pass→pass | 19,327 | 20,553 | +6% | 1 | 1 | 0% | 2,537 | 3,290 | +30% | 0 | 0 | — |
case-12 | pass→pass | 28,112 | 23,495 | -16% | 1 | 1 | 0% | 3,814 | 4,203 | +10% | 0 | 0 | — |
case-13 | pass→pass | 19,930 | 25,047 | +26% | 1 | 1 | 0% | 2,804 | 3,433 | +22% | 0 | 0 | — |
case-14 | pass→pass | 18,438 | 21,586 | +17% | 1 | 1 | 0% | 2,783 | 3,342 | +20% | 0 | 0 | — |
case-15 | pass→pass | 33,778 | 37,891 | +12% | 1 | 1 | 0% | 3,261 | 3,926 | +20% | 0 | 0 | — |
case-16 | pass→pass | 23,374 | 14,713 | -37% | 1 | 1 | 0% | 2,565 | 2,706 | +5% | 0 | 0 | — |
case-17 | pass→pass | 23,200 | 30,075 | +30% | 1 | 1 | 0% | 2,892 | 3,134 | +8% | 0 | 0 | — |
case-18 | pass→pass | 21,947 | 27,892 | +27% | 1 | 1 | 0% | 3,282 | 4,009 | +22% | 0 | 0 | — |
case-19 | pass→pass | 20,374 | 15,439 | -24% | 1 | 1 | 0% | 2,643 | 2,919 | +10% | 0 | 0 | — |
case-20 | pass→pass | 32,853 | 94,137 | +187% | 1 | 1 | 0% | 6,480 | 7,252 | +12% | 0 | 0 | — |
case-21 | pass→pass | 15,272 | 20,112 | +32% | 1 | 1 | 0% | 3,464 | 3,989 | +15% | 0 | 0 | — |
case-22 | pass→pass | 27,965 | 28,347 | +1% | 1 | 1 | 0% | 5,260 | 6,411 | +22% | 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. The headline lift of +5 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.