Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股尾盘策略/收盘效应分析。当用户说"尾盘"、"收盘效应"、"尾盘拉升"、"尾盘跳水"、"14:57"、"收盘集合竞价"、"尾盘策略"、"最后半小时"时触发。基于 cn-stock-data 获取数据,分析尾盘异动、收盘效应、尾盘交易策略。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-closing-strategy/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 49% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 38% | 0% |
| case-06 | ✓→✓ | = Same ✓ | 30% | 0% |
通过 cn-stock-data skill 获取数据:
# [标的/市场] 尾盘分析报告
## 一、尾盘异动
| 指标 | 数值 | 信号 |
|------|------|------|
## 二、收盘效应
[当前是否存在特殊效应]
## 三、次日预测
[基于尾盘信号的次日走势预判]
## 四、策略建议## [标的] 尾盘速览
- 尾盘30分钟涨 +0.8%,占全天涨幅65%
- 尾盘放量,最后30分钟量占28%
- 非月末/交割日,排除特殊效应
- 次日高开概率 58%,但追高需谨慎参考 references/closing-strategy-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-04 | pass→pass | 23,338 | 28,603 | +23% | 1 | 1 | 0% | 3,259 | 4,860 | +49% | 0 | 0 | — |
case-01 | fail→fail | 16,776 | 25,168 | +50% | 1 | 1 | 0% | 2,620 | 3,532 | +35% | 0 | 0 | — |
case-02 | fail→fail | 26,973 | 22,724 | -16% | 1 | 1 | 0% | 3,864 | 4,461 | +15% | 0 | 0 | — |
case-03 | fail→fail | 20,370 | 24,775 | +22% | 1 | 1 | 0% | 3,171 | 4,700 | +48% | 0 | 0 | — |
case-05 | pass→pass | 18,373 | 23,202 | +26% | 1 | 1 | 0% | 2,898 | 3,989 | +38% | 0 | 0 | — |
case-06 | pass→pass | 22,222 | 24,665 | +11% | 1 | 1 | 0% | 3,159 | 4,100 | +30% | 0 | 0 | — |
case-07 | fail→pass | 21,684 | 16,303 | -25% | 1 | 1 | 0% | 2,802 | 3,284 | +17% | 0 | 0 | — |
case-08 | fail→pass | 14,754 | 14,357 | -3% | 1 | 1 | 0% | 2,390 | 3,054 | +28% | 0 | 0 | — |
case-09 | fail→fail | 5,574 | 10,816 | +94% | 1 | 1 | 0% | 1,032 | 2,307 | +124% | 0 | 0 | — |
case-10 | pass→pass | 11,294 | 15,770 | +40% | 1 | 1 | 0% | 1,859 | 3,031 | +63% | 0 | 0 | — |
case-11 | pass→pass | 16,328 | 17,295 | +6% | 1 | 1 | 0% | 2,674 | 3,752 | +40% | 0 | 0 | — |
case-12 | pass→pass | 19,001 | 17,209 | -9% | 1 | 1 | 0% | 2,587 | 3,343 | +29% | 0 | 0 | — |
case-13 | pass→pass | 20,849 | 22,886 | +10% | 1 | 1 | 0% | 2,995 | 4,042 | +35% | 0 | 0 | — |
case-18 | fail→fail | 15,416 | 7,072 | -54% | 1 | 1 | 0% | 2,486 | 1,879 | -24% | 0 | 0 | — |
case-14 | pass→pass | 18,375 | 15,468 | -16% | 1 | 1 | 0% | 2,439 | 3,160 | +30% | 0 | 0 | — |
case-15 | pass→pass | 21,024 | 18,149 | -14% | 1 | 1 | 0% | 3,070 | 3,501 | +14% | 0 | 0 | — |
case-16 | fail→fail | 17,556 | 18,151 | +3% | 1 | 1 | 0% | 2,939 | 3,354 | +14% | 0 | 0 | — |
case-17 | fail→fail | 11,267 | 11,458 | +2% | 1 | 1 | 0% | 1,827 | 2,481 | +36% | 0 | 0 | — |
case-19 | pass→pass | 21,523 | 22,489 | +4% | 1 | 1 | 0% | 3,275 | 4,234 | +29% | 0 | 0 | — |
case-20 | pass→pass | 26,402 | 26,453 | +0% | 1 | 1 | 0% | 3,592 | 4,427 | +23% | 0 | 0 | — |
case-21 | pass→pass | 22,080 | 26,036 | +18% | 1 | 1 | 0% | 3,083 | 4,118 | +34% | 0 | 0 | — |
case-22 | pass→pass | 23,489 | 22,784 | -3% | 1 | 1 | 0% | 3,222 | 4,247 | +32% | 0 | 0 | — |
case-23 | pass→pass | 20,535 | 20,842 | +1% | 1 | 1 | 0% | 2,803 | 3,718 | +33% | 0 | 0 | — |
case-24 | pass→pass | 22,937 | 24,009 | +5% | 1 | 1 | 0% | 3,322 | 4,268 | +28% | 0 | 0 | — |
case-25 | pass→pass | 23,037 | 23,321 | +1% | 1 | 1 | 0% | 3,376 | 4,334 | +28% | 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. 25 cases were attempted. The headline lift of +8 percentage points is the difference between those two pass rates over the 25 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 | +9% |
Other measured skills in the registry, with their headline benchmark lift.