Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在银行风险管理场景下，围绕保证担保人的资质变化、代偿风险和担保覆盖度进行持续监测与预警时使用本技能。适合输出担保风险摘要、优先级判断、处置建议和升级路径。
.claude/skills/aifinlab-bank-t202-risk-management-guarantee-risk-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 68% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 6% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
本技能面向银行风险管理场景，聚焦保证担保人资质、代偿能力和担保覆盖度的持续监测与风险识别。目标是将规则命中、担保人画像变化、代偿历史与风险信号转成可执行的排查与处置动作，便于风控、授信管理和预警运营团队快速响应。
scripts/guarantee_risk_assess.py：用于担保风险信号分级、覆盖度校验与报告摘要生成| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,945 | 42,125 | +56% | 1 | 1 | 0% | 4,044 | 4,557 | +13% | 0 | 0 | — |
case-02 | fail→fail | 27,892 | 25,975 | -7% | 1 | 1 | 0% | 4,277 | 4,288 | +0% | 0 | 0 | — |
case-03 | fail→fail | 31,233 | 25,941 | -17% | 1 | 1 | 0% | 3,913 | 4,886 | +25% | 0 | 0 | — |
case-04 | pass→pass | 19,124 | 22,323 | +17% | 1 | 1 | 0% | 2,144 | 3,611 | +68% | 0 | 0 | — |
case-05 | pass→pass | 20,160 | 16,783 | -17% | 1 | 1 | 0% | 2,745 | 2,921 | +6% | 0 | 0 | — |
case-06 | pass→pass | 19,442 | 15,560 | -20% | 1 | 1 | 0% | 2,398 | 3,124 | +30% | 0 | 0 | — |
case-07 | fail→fail | 21,905 | 19,055 | -13% | 1 | 1 | 0% | 3,058 | 3,673 | +20% | 0 | 0 | — |
case-08 | fail→pass | 24,091 | 29,624 | +23% | 1 | 1 | 0% | 3,267 | 4,254 | +30% | 0 | 0 | — |
case-09 | fail→fail | 23,508 | 22,117 | -6% | 1 | 1 | 0% | 3,604 | 3,985 | +11% | 0 | 0 | — |
case-10 | fail→fail | 22,425 | 23,395 | +4% | 1 | 1 | 0% | 2,580 | 3,592 | +39% | 0 | 0 | — |
case-11 | fail→fail | 26,909 | 22,419 | -17% | 1 | 1 | 0% | 3,365 | 4,005 | +19% | 0 | 0 | — |
case-12 | pass→pass | 22,548 | 21,496 | -5% | 1 | 1 | 0% | 3,131 | 3,510 | +12% | 0 | 0 | — |
case-13 | fail→fail | 22,098 | 22,413 | +1% | 1 | 1 | 0% | 3,114 | 3,685 | +18% | 0 | 0 | — |
case-14 | fail→fail | 21,235 | 25,567 | +20% | 1 | 1 | 0% | 3,149 | 4,674 | +48% | 0 | 0 | — |
case-15 | fail→fail | 29,104 | 25,361 | -13% | 1 | 1 | 0% | 3,891 | 4,444 | +14% | 0 | 0 | — |
case-16 | fail→fail | 25,674 | 27,316 | +6% | 1 | 1 | 0% | 3,667 | 4,278 | +17% | 0 | 0 | — |
case-17 | fail→fail | 23,601 | 22,742 | -4% | 1 | 1 | 0% | 3,031 | 3,954 | +30% | 0 | 0 | — |
case-18 | fail→pass | 25,148 | 24,909 | -1% | 1 | 1 | 0% | 3,613 | 4,253 | +18% | 0 | 0 | — |
case-19 | fail→fail | 23,480 | 25,982 | +11% | 1 | 1 | 0% | 3,320 | 4,205 | +27% | 0 | 0 | — |
case-20 | fail→fail | 23,193 | 22,669 | -2% | 1 | 1 | 0% | 2,978 | 3,536 | +19% | 0 | 0 | — |
case-21 | fail→fail | 23,126 | 28,509 | +23% | 1 | 1 | 0% | 2,993 | 3,945 | +32% | 0 | 0 | — |
case-22 | fail→fail | 22,171 | 23,989 | +8% | 1 | 1 | 0% | 2,989 | 4,047 | +35% | 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 +9 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.
Other measured skills in the registry, with their headline benchmark lift.