Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股动态对冲/Delta中性策略。当用户说"动态对冲"、"Delta对冲"、"Delta中性"、"dynamic hedging"、"对冲调整"、"对冲比例"时触发。基于 cn-stock-data 获取数据，设计动态对冲方案。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-dynamic-hedging/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-23 | ✓→✗ | ▼ Worse | 67% | 0% |
| case-14 | ✓→✓ | ＝ Same ✓ | -19% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -4% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 17% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 41% | 0% |
通过 cn-stock-data skill 获取数据：
# 动态对冲方案报告
## 一、对冲需求
| 指标 | 数值 |
|------|------|
| 组合Delta | +0.85 |
## 二、对冲方案
[工具选择、对冲比例]
## 三、调整规则
[触发条件、调整频率]
## 四、成本与效果## 动态对冲速览
- 组合Delta +0.85，需对冲
- 方案：卖出IF期货3手
- 对冲后Delta降至+0.05
- 年化对冲成本约2%(期货贴水)参考 references/dynamic-hedging-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-14 | pass→pass | 12,674 | 6,876 | -46% | 1 | 1 | 0% | 2,066 | 1,676 | -19% | 0 | 0 | — |
case-01 | pass→pass | 41,233 | 42,806 | +4% | 1 | 1 | 0% | 5,711 | 5,498 | -4% | 0 | 0 | — |
case-02 | pass→pass | 15,635 | 15,769 | +1% | 1 | 1 | 0% | 2,569 | 2,994 | +17% | 0 | 0 | — |
case-03 | pass→pass | 29,533 | 40,794 | +38% | 1 | 1 | 0% | 4,715 | 6,634 | +41% | 0 | 0 | — |
case-04 | pass→pass | 5,860 | 4,834 | -18% | 1 | 1 | 0% | 1,124 | 1,566 | +39% | 0 | 0 | — |
case-05 | pass→pass | 6,079 | 9,145 | +50% | 1 | 1 | 0% | 1,274 | 2,289 | +80% | 0 | 0 | — |
case-06 | pass→pass | 4,306 | 6,688 | +55% | 1 | 1 | 0% | 962 | 1,670 | +74% | 0 | 0 | — |
case-07 | pass→pass | 21,677 | 21,760 | +0% | 1 | 1 | 0% | 3,125 | 4,469 | +43% | 0 | 0 | — |
case-08 | pass→pass | 9,603 | 10,783 | +12% | 1 | 1 | 0% | 1,607 | 2,278 | +42% | 0 | 0 | — |
case-20 | pass→pass | 19,237 | 20,523 | +7% | 1 | 1 | 0% | 2,846 | 3,630 | +28% | 0 | 0 | — |
case-09 | pass→pass | 8,517 | 7,860 | -8% | 1 | 1 | 0% | 1,294 | 2,068 | +60% | 0 | 0 | — |
case-10 | pass→pass | 11,869 | 12,622 | +6% | 1 | 1 | 0% | 1,875 | 2,545 | +36% | 0 | 0 | — |
case-11 | pass→pass | 16,016 | 15,703 | -2% | 1 | 1 | 0% | 2,385 | 3,019 | +27% | 0 | 0 | — |
case-12 | pass→pass | 17,168 | 14,895 | -13% | 1 | 1 | 0% | 2,356 | 2,727 | +16% | 0 | 0 | — |
case-13 | pass→pass | 13,500 | 17,808 | +32% | 1 | 1 | 0% | 2,356 | 3,773 | +60% | 0 | 0 | — |
case-15 | pass→pass | 18,649 | 19,957 | +7% | 1 | 1 | 0% | 3,290 | 4,137 | +26% | 0 | 0 | — |
case-16 | pass→pass | 15,844 | 16,363 | +3% | 1 | 1 | 0% | 2,386 | 3,207 | +34% | 0 | 0 | — |
case-17 | pass→pass | 20,252 | 18,107 | -11% | 1 | 1 | 0% | 2,714 | 3,384 | +25% | 0 | 0 | — |
case-18 | pass→pass | 15,439 | 15,736 | +2% | 1 | 1 | 0% | 2,862 | 3,437 | +20% | 0 | 0 | — |
case-19 | pass→pass | 20,612 | 22,369 | +9% | 1 | 1 | 0% | 3,268 | 4,331 | +33% | 0 | 0 | — |
case-21 | pass→pass | 19,654 | 23,859 | +21% | 1 | 1 | 0% | 3,460 | 4,810 | +39% | 0 | 0 | — |
case-22 | pass→pass | 21,866 | 36,884 | +69% | 1 | 1 | 0% | 3,714 | 6,261 | +69% | 0 | 0 | — |
case-23 | pass→fail | 25,645 | 30,668 | +20% | 1 | 1 | 0% | 3,169 | 5,289 | +67% | 0 | 0 | — |
case-24 | pass→pass | 15,763 | 19,643 | +25% | 1 | 1 | 0% | 3,036 | 3,781 | +25% | 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. 24 cases were attempted. The headline lift of -100 percentage points is the difference between those two pass rates over the 24 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/28/2026 | +9% |
Other measured skills in the registry, with their headline benchmark lift.