Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股日历价差/跨期策略。当用户说"日历价差"、"calendar spread"、"跨期"、"近远月价差"、"时间价差"、"跨期套利"时触发。基于 cn-stock-data 获取数据，分析期权/期货日历价差策略。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-calendar-spread/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-20 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-05 | ✓→✗ | ▼ Worse | 61% | 0% |
通过 cn-stock-data skill 获取数据：
# 日历价差策略报告
## 一、价差分析
| 组合 | 近月 | 远月 | 价差 | 分位 |
|------|------|------|------|------|
## 二、策略方案
[具体合约、Greeks]
## 三、情景分析
[标的±5%/IV±5%盈亏]
## 四、展期计划## 日历价差速览
- 近月ATM Call 0.15，远月 0.28
- 价差 0.13，历史P40
- 建议：买入日历价差(IV偏低)
- 最大亏损：净支出0.13参考 references/calendar-spread-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-18 | pass→pass | 19,755 | 14,019 | -29% | 1 | 1 | 0% | 3,463 | 3,121 | -10% | 0 | 0 | — |
case-12 | pass→pass | 24,519 | 23,174 | -5% | 1 | 1 | 0% | 3,475 | 4,019 | +16% | 0 | 0 | — |
case-01 | fail→fail | 30,085 | 49,196 | +64% | 1 | 1 | 0% | 4,989 | 6,192 | +24% | 0 | 0 | — |
case-02 | fail→pass | 13,098 | 12,250 | -6% | 1 | 1 | 0% | 1,900 | 2,857 | +50% | 0 | 0 | — |
case-17 | pass→pass | 17,310 | 15,137 | -13% | 1 | 1 | 0% | 2,833 | 3,387 | +20% | 0 | 0 | — |
case-03 | fail→fail | 25,628 | 26,018 | +2% | 1 | 1 | 0% | 4,014 | 6,037 | +50% | 0 | 0 | — |
case-04 | pass→pass | 16,961 | 16,871 | -1% | 1 | 1 | 0% | 2,436 | 3,118 | +28% | 0 | 0 | — |
case-05 | pass→fail | 20,666 | 22,136 | +7% | 1 | 1 | 0% | 2,445 | 3,934 | +61% | 0 | 0 | — |
case-06 | fail→pass | 21,079 | 19,731 | -6% | 1 | 1 | 0% | 3,074 | 3,794 | +23% | 0 | 0 | — |
case-07 | fail→fail | 14,341 | 17,318 | +21% | 1 | 1 | 0% | 2,065 | 3,170 | +54% | 0 | 0 | — |
case-08 | pass→pass | 10,640 | 15,113 | +42% | 1 | 1 | 0% | 1,746 | 2,898 | +66% | 0 | 0 | — |
case-09 | pass→pass | 14,335 | 12,405 | -13% | 1 | 1 | 0% | 2,250 | 2,682 | +19% | 0 | 0 | — |
case-10 | pass→pass | 12,941 | 13,994 | +8% | 1 | 1 | 0% | 2,230 | 2,625 | +18% | 0 | 0 | — |
case-11 | pass→pass | 9,999 | 10,803 | +8% | 1 | 1 | 0% | 1,695 | 2,503 | +48% | 0 | 0 | — |
case-13 | fail→pass | 19,257 | 7,012 | -64% | 1 | 1 | 0% | 3,284 | 1,935 | -41% | 0 | 0 | — |
case-14 | fail→fail | 13,938 | 18,986 | +36% | 1 | 1 | 0% | 2,423 | 3,728 | +54% | 0 | 0 | — |
case-15 | fail→fail | 10,432 | 13,966 | +34% | 1 | 1 | 0% | 1,776 | 2,624 | +48% | 0 | 0 | — |
case-16 | pass→pass | 13,373 | 14,735 | +10% | 1 | 1 | 0% | 2,368 | 3,119 | +32% | 0 | 0 | — |
case-19 | pass→pass | 17,471 | 23,687 | +36% | 1 | 1 | 0% | 2,859 | 4,425 | +55% | 0 | 0 | — |
case-20 | fail→pass | 20,512 | 4,192 | -80% | 1 | 1 | 0% | 3,000 | 1,416 | -53% | 0 | 0 | — |
case-21 | pass→pass | 13,128 | 15,897 | +21% | 1 | 1 | 0% | 1,804 | 2,804 | +55% | 0 | 0 | — |
case-22 | pass→pass | 20,277 | 20,310 | +0% | 1 | 1 | 0% | 3,325 | 4,129 | +24% | 0 | 0 | — |
case-23 | pass→pass | 18,179 | 22,108 | +22% | 1 | 1 | 0% | 3,327 | 4,285 | +29% | 0 | 0 | — |
case-24 | pass→pass | 19,082 | 19,742 | +3% | 1 | 1 | 0% | 3,376 | 4,158 | +23% | 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 +13 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.
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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/28/2026 | +9% |
Other measured skills in the registry, with their headline benchmark lift.