Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当需要对风险指标或异常现象进行成因拆解、驱动项解释并输出业务提示时使用本技能。
.claude/skills/aifinlab-bank-t215-risk-management-risk-explanation-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -72% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-23 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 20% | 0% |
用于对风险指标异常、趋势变化或事件现象进行解释性拆解，形成结构化诊断结论、驱动项说明、影响提示与后续追问方向。
scripts/risk_explanation_builder.py：生成风险解释与驱动项拆解草稿。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 30,153 | 24,363 | -19% | 1 | 1 | 0% | 3,768 | 3,990 | +6% | 0 | 0 | — |
case-02 | pass→pass | 20,247 | 18,109 | -11% | 1 | 1 | 0% | 2,508 | 3,020 | +20% | 0 | 0 | — |
case-03 | pass→pass | 21,662 | 22,715 | +5% | 1 | 1 | 0% | 3,024 | 3,323 | +10% | 0 | 0 | — |
case-04 | pass→pass | 21,993 | 17,382 | -21% | 1 | 1 | 0% | 2,773 | 2,876 | +4% | 0 | 0 | — |
case-05 | pass→pass | 18,443 | 22,512 | +22% | 1 | 1 | 0% | 2,869 | 3,138 | +9% | 0 | 0 | — |
case-06 | fail→fail | 20,367 | 23,521 | +15% | 1 | 1 | 0% | 2,625 | 3,476 | +32% | 0 | 0 | — |
case-07 | pass→pass | 16,681 | 14,665 | -12% | 1 | 1 | 0% | 2,406 | 2,450 | +2% | 0 | 0 | — |
case-08 | pass→pass | 26,592 | 27,202 | +2% | 1 | 1 | 0% | 3,392 | 3,369 | -1% | 0 | 0 | — |
case-09 | pass→pass | 22,826 | 23,121 | +1% | 1 | 1 | 0% | 3,265 | 3,227 | -1% | 0 | 0 | — |
case-10 | pass→pass | 19,695 | 19,232 | -2% | 1 | 1 | 0% | 2,721 | 3,119 | +15% | 0 | 0 | — |
case-11 | pass→pass | 11,076 | 15,298 | +38% | 1 | 1 | 0% | 1,481 | 2,315 | +56% | 0 | 0 | — |
case-12 | pass→pass | 17,104 | 14,177 | -17% | 1 | 1 | 0% | 2,205 | 2,352 | +7% | 0 | 0 | — |
case-13 | pass→pass | 20,900 | 20,259 | -3% | 1 | 1 | 0% | 2,661 | 2,852 | +7% | 0 | 0 | — |
case-14 | pass→pass | 21,830 | 22,045 | +1% | 1 | 1 | 0% | 2,965 | 3,515 | +19% | 0 | 0 | — |
case-15 | pass→pass | 22,162 | 31,922 | +44% | 1 | 1 | 0% | 2,766 | 3,089 | +12% | 0 | 0 | — |
case-16 | fail→fail | 26,713 | 25,229 | -6% | 1 | 1 | 0% | 3,353 | 3,581 | +7% | 0 | 0 | — |
case-17 | pass→pass | 18,608 | 18,061 | -3% | 1 | 1 | 0% | 2,595 | 2,890 | +11% | 0 | 0 | — |
case-18 | pass→pass | 22,031 | 24,327 | +10% | 1 | 1 | 0% | 3,514 | 3,609 | +3% | 0 | 0 | — |
case-19 | fail→pass | 18,592 | 2,867 | -85% | 1 | 1 | 0% | 2,508 | 706 | -72% | 0 | 0 | — |
case-20 | fail→fail | 19,679 | 21,656 | +10% | 1 | 1 | 0% | 2,668 | 2,997 | +12% | 0 | 0 | — |
case-21 | pass→pass | 23,475 | 17,949 | -24% | 1 | 1 | 0% | 2,978 | 2,888 | -3% | 0 | 0 | — |
case-22 | fail→pass | 19,678 | 25,006 | +27% | 1 | 1 | 0% | 2,774 | 3,791 | +37% | 0 | 0 | — |
case-23 | fail→pass | 14,113 | 15,659 | +11% | 1 | 1 | 0% | 1,883 | 2,241 | +19% | 0 | 0 | — |
case-24 | pass→pass | 20,873 | 23,804 | +14% | 1 | 1 | 0% | 2,645 | 3,767 | +42% | 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 +17 percentage points is the difference between those two pass rates over the 24 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.
Other measured skills in the registry, with their headline benchmark lift.