Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股交易成本分析/TCA。当用户说"交易成本"、"TCA"、"transaction cost"、"成本分析"、"佣金"、"印花税"、"交易费用"时触发。基于 cn-stock-data 获取数据，进行全面的交易成本分析。支持 formal/brief 两种输出风格。
.claude/skills/aifinlab-a-share-transaction-cost/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-20 | ✓→✗ | ▼ Worse | 26% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 38% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | -1% | 0% |
通过 cn-stock-data skill 获取数据：
# 交易成本分析(TCA)报告
## 一、成本汇总
| 成本类型 | 金额 | 占比 |
|---------|------|------|
## 二、IS分解
[各成本成分明细]
## 三、对标分析
[与基准对比]
## 四、优化建议## TCA速览
- 总交易成本 0.18%/次
- 显性0.06% + 隐性0.12%
- 冲击成本占隐性成本65%
- 建议：使用VWAP算法可降低冲击30%参考 references/transaction-cost-guide.md 获取详细方法论与 A股实证研究。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 9,322 | 9,553 | +2% | 1 | 1 | 0% | 1,641 | 2,265 | +38% | 0 | 0 | — |
case-01 | fail→fail | 38,174 | 34,421 | -10% | 1 | 1 | 0% | 5,256 | 6,175 | +17% | 0 | 0 | — |
case-02 | fail→fail | 14,033 | 11,381 | -19% | 1 | 1 | 0% | 2,091 | 2,435 | +16% | 0 | 0 | — |
case-03 | fail→fail | 28,542 | 23,854 | -16% | 1 | 1 | 0% | 4,480 | 4,706 | +5% | 0 | 0 | — |
case-05 | fail→fail | 11,699 | 15,737 | +35% | 1 | 1 | 0% | 2,262 | 3,427 | +52% | 0 | 0 | — |
case-06 | fail→fail | 20,043 | 19,236 | -4% | 1 | 1 | 0% | 3,095 | 3,599 | +16% | 0 | 0 | — |
case-07 | pass→pass | 13,087 | 8,299 | -37% | 1 | 1 | 0% | 1,880 | 1,869 | -1% | 0 | 0 | — |
case-08 | fail→fail | 35,680 | 24,819 | -30% | 1 | 1 | 0% | 5,118 | 4,966 | -3% | 0 | 0 | — |
case-09 | fail→fail | 13,748 | 11,511 | -16% | 1 | 1 | 0% | 2,283 | 2,193 | -4% | 0 | 0 | — |
case-10 | fail→pass | 19,066 | 18,340 | -4% | 1 | 1 | 0% | 2,851 | 3,035 | +6% | 0 | 0 | — |
case-11 | fail→pass | 13,779 | 4,478 | -68% | 1 | 1 | 0% | 1,902 | 1,161 | -39% | 0 | 0 | — |
case-12 | pass→pass | 19,317 | 15,467 | -20% | 1 | 1 | 0% | 2,770 | 3,047 | +10% | 0 | 0 | — |
case-13 | pass→pass | 19,517 | 16,072 | -18% | 1 | 1 | 0% | 3,343 | 3,284 | -2% | 0 | 0 | — |
case-14 | pass→pass | 13,837 | 11,606 | -16% | 1 | 1 | 0% | 1,961 | 2,330 | +19% | 0 | 0 | — |
case-15 | fail→fail | 12,336 | 2,891 | -77% | 1 | 1 | 0% | 2,109 | 1,081 | -49% | 0 | 0 | — |
case-16 | pass→pass | 26,902 | 24,307 | -10% | 1 | 1 | 0% | 3,364 | 3,904 | +16% | 0 | 0 | — |
case-17 | pass→pass | 19,990 | 18,671 | -7% | 1 | 1 | 0% | 3,124 | 3,505 | +12% | 0 | 0 | — |
case-18 | pass→pass | 19,853 | 19,713 | -1% | 1 | 1 | 0% | 2,822 | 3,671 | +30% | 0 | 0 | — |
case-19 | pass→pass | 17,136 | 17,738 | +4% | 1 | 1 | 0% | 2,696 | 3,053 | +13% | 0 | 0 | — |
case-20 | pass→fail | 25,507 | 35,166 | +38% | 1 | 1 | 0% | 4,237 | 5,347 | +26% | 0 | 0 | — |
case-21 | pass→pass | 10,751 | 13,524 | +26% | 1 | 1 | 0% | 2,196 | 2,617 | +19% | 0 | 0 | — |
case-22 | pass→pass | 21,737 | 34,186 | +57% | 1 | 1 | 0% | 3,334 | 5,775 | +73% | 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. 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.
Other measured skills in the registry, with their headline benchmark lift.