Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当需要将多源风险信号按客户/集团/产品维度聚合，并输出优先级、处置建议和跟进清单时使用本技能，适合风控、预警运营和审查支持场景。
.claude/skills/aifinlab-bank-t210-risk-management-risk-aggregation-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 10% | 0% |
用于把来自交易、授信、名单、押品、舆情等多源风险信号进行聚合归并，形成可执行的风险视图：包括信号摘要、优先级、影响范围、建议动作与升级条件。
scripts/risk_aggregation_report.py：读取信号与对象信息，生成聚合报告草稿与优先级建议。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 36,752 | 33,513 | -9% | 1 | 1 | 0% | 4,711 | 4,211 | -11% | 0 | 0 | — |
case-02 | fail→fail | 23,001 | 24,202 | +5% | 1 | 1 | 0% | 3,620 | 3,666 | +1% | 0 | 0 | — |
case-03 | fail→pass | 23,932 | 19,901 | -17% | 1 | 1 | 0% | 2,947 | 3,190 | +8% | 0 | 0 | — |
case-04 | fail→pass | 17,489 | 24,159 | +38% | 1 | 1 | 0% | 2,280 | 3,179 | +39% | 0 | 0 | — |
case-05 | fail→pass | 14,481 | 9,490 | -34% | 1 | 1 | 0% | 1,838 | 1,987 | +8% | 0 | 0 | — |
case-06 | fail→pass | 17,396 | 15,529 | -11% | 1 | 1 | 0% | 2,575 | 2,825 | +10% | 0 | 0 | — |
case-07 | fail→pass | 18,686 | 13,797 | -26% | 1 | 1 | 0% | 2,419 | 2,669 | +10% | 0 | 0 | — |
case-08 | fail→fail | 18,011 | 22,500 | +25% | 1 | 1 | 0% | 2,097 | 3,975 | +90% | 0 | 0 | — |
case-09 | pass→pass | 17,200 | 17,985 | +5% | 1 | 1 | 0% | 2,222 | 3,177 | +43% | 0 | 0 | — |
case-10 | fail→pass | 18,908 | 18,786 | -1% | 1 | 1 | 0% | 2,192 | 2,970 | +35% | 0 | 0 | — |
case-11 | fail→pass | 27,417 | 24,475 | -11% | 1 | 1 | 0% | 3,392 | 4,378 | +29% | 0 | 0 | — |
case-12 | fail→fail | 16,404 | 18,916 | +15% | 1 | 1 | 0% | 2,184 | 3,525 | +61% | 0 | 0 | — |
case-13 | fail→fail | 18,759 | 21,327 | +14% | 1 | 1 | 0% | 2,361 | 3,135 | +33% | 0 | 0 | — |
case-14 | pass→pass | 22,634 | 16,469 | -27% | 1 | 1 | 0% | 3,098 | 2,870 | -7% | 0 | 0 | — |
case-15 | pass→pass | 15,652 | 20,380 | +30% | 1 | 1 | 0% | 2,249 | 3,132 | +39% | 0 | 0 | — |
case-16 | pass→pass | 23,290 | 22,232 | -5% | 1 | 1 | 0% | 3,023 | 3,454 | +14% | 0 | 0 | — |
case-17 | pass→pass | 20,678 | 26,021 | +26% | 1 | 1 | 0% | 2,730 | 2,957 | +8% | 0 | 0 | — |
case-18 | fail→pass | 24,550 | 14,390 | -41% | 1 | 1 | 0% | 2,830 | 2,789 | -1% | 0 | 0 | — |
case-19 | pass→pass | 27,846 | 21,266 | -24% | 1 | 1 | 0% | 2,696 | 3,200 | +19% | 0 | 0 | — |
case-20 | fail→fail | 42,431 | 24,538 | -42% | 1 | 1 | 0% | 3,126 | 3,736 | +20% | 0 | 0 | — |
case-21 | pass→pass | 19,974 | 17,440 | -13% | 1 | 1 | 0% | 2,643 | 3,039 | +15% | 0 | 0 | — |
case-22 | pass→pass | 19,468 | 21,553 | +11% | 1 | 1 | 0% | 2,527 | 3,288 | +30% | 0 | 0 | — |
case-23 | fail→pass | 21,041 | 22,851 | +9% | 1 | 1 | 0% | 3,166 | 3,986 | +26% | 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. 23 cases were attempted. The headline lift of +39 percentage points is the difference between those two pass rates over the 23 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.