Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股日内波动率/已实现波动率分析。当用户说"日内波动率"、"已实现波动率"、"realized volatility"、"RV"、"盘中波动"、"日内波动"、"高频波动率"时触发。基于 cn-stock-data 获取数据,计算已实现波动率、分析日内波动率模式、预测未来波动。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-intraday-vol/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 3% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 6% | 0% |
通过 cn-stock-data skill 获取数据:
# [标的] 日内波动率分析报告
## 一、已实现波动率
| 指标 | 数值 | 分位数 |
|------|------|--------|
| 日RV | 1.8% | P65 |
## 二、日内模式
[U型曲线、异常时段]
## 三、波动率预测
[HAR-RV预测、状态判断]
## 四、交易建议
[基于波动率的仓位与策略建议]## [标的] 波动率速览
- 日RV 1.8% (P65),波动率中等偏高
- 开盘30分钟贡献全天波动率42%
- HAR-RV预测明日RV 1.6%,略降
- 当前处于正常波动状态参考 references/intraday-vol-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 | 42,667 | 33,081 | -22% | 1 | 1 | 0% | 6,175 | 6,143 | -1% | 0 | 0 | — |
case-02 | fail→pass | 20,205 | 14,589 | -28% | 1 | 1 | 0% | 2,851 | 2,420 | -15% | 0 | 0 | — |
case-03 | fail→fail | 30,746 | 46,610 | +52% | 1 | 1 | 0% | 4,475 | 8,853 | +98% | 0 | 0 | — |
case-04 | pass→pass | 39,852 | 16,208 | -59% | 1 | 1 | 0% | 3,279 | 3,363 | +3% | 0 | 0 | — |
case-05 | pass→pass | 18,385 | 15,300 | -17% | 1 | 1 | 0% | 3,010 | 3,185 | +6% | 0 | 0 | — |
case-06 | pass→pass | 21,100 | 19,718 | -7% | 1 | 1 | 0% | 3,356 | 3,760 | +12% | 0 | 0 | — |
case-07 | pass→pass | 19,125 | 17,131 | -10% | 1 | 1 | 0% | 3,042 | 3,627 | +19% | 0 | 0 | — |
case-08 | pass→pass | 17,907 | 14,842 | -17% | 1 | 1 | 0% | 2,289 | 2,696 | +18% | 0 | 0 | — |
case-09 | pass→pass | 9,548 | 11,740 | +23% | 1 | 1 | 0% | 1,526 | 2,426 | +59% | 0 | 0 | — |
case-10 | pass→pass | 17,920 | 17,827 | -1% | 1 | 1 | 0% | 2,822 | 3,575 | +27% | 0 | 0 | — |
case-11 | pass→pass | 22,445 | 17,723 | -21% | 1 | 1 | 0% | 2,929 | 3,304 | +13% | 0 | 0 | — |
case-12 | pass→pass | 18,133 | 17,926 | -1% | 1 | 1 | 0% | 2,757 | 3,374 | +22% | 0 | 0 | — |
case-13 | pass→pass | 26,080 | 20,158 | -23% | 1 | 1 | 0% | 3,131 | 3,644 | +16% | 0 | 0 | — |
case-14 | pass→pass | 15,980 | 16,959 | +6% | 1 | 1 | 0% | 2,116 | 3,342 | +58% | 0 | 0 | — |
case-15 | pass→pass | 7,608 | 10,408 | +37% | 1 | 1 | 0% | 1,209 | 2,212 | +83% | 0 | 0 | — |
case-16 | fail→fail | 25,547 | 27,992 | +10% | 1 | 1 | 0% | 3,978 | 5,031 | +26% | 0 | 0 | — |
case-17 | fail→pass | 16,628 | 10,640 | -36% | 1 | 1 | 0% | 2,308 | 2,123 | -8% | 0 | 0 | — |
case-18 | fail→pass | 17,929 | 3,120 | -83% | 1 | 1 | 0% | 2,746 | 1,199 | -56% | 0 | 0 | — |
case-19 | pass→pass | 23,970 | 20,781 | -13% | 1 | 1 | 0% | 3,320 | 3,725 | +12% | 0 | 0 | — |
case-20 | pass→pass | 28,843 | 35,478 | +23% | 1 | 1 | 0% | 5,719 | 8,208 | +44% | 0 | 0 | — |
case-21 | pass→pass | 28,222 | 31,438 | +11% | 1 | 1 | 0% | 4,057 | 5,853 | +44% | 0 | 0 | — |
case-22 | pass→pass | 20,365 | 25,316 | +24% | 1 | 1 | 0% | 3,287 | 4,728 | +44% | 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 +14 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/28/2026 | 0% |
Other measured skills in the registry, with their headline benchmark lift.