Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在银行零售金融场景下，围绕关怀推荐进行适配、配置、推荐、方案设计或备选方案比较时使用本技能。适合输出主方案、备选方案、匹配逻辑、风险提示和沟通要点。
.claude/skills/aifinlab-bank-t174-retail-finance-care-recommendation-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-24 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 156% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | 3% | 0% |
在零售金融场景下，为客户生日或纪念日等节点输出关怀推荐方案，明确触达方式、关怀力度、主方案与备选方案，并提示合规边界。
batch_id、time_windowcustomers[]（建议字段 customer_id、name、segment、value_segment、upcoming_event、event_date、preferred_channel、consent_marketing、risk_flags）脚本入口：scripts/run_skill.py，调用 shared/retail_service_skill_engine.py 的 t174 场景。
bashpython scripts/run_skill.py --input input.json --format markdown python scripts/run_skill.py --input input.json --format json
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | pass→pass | 8,041 | 3,073 | -62% | 1 | 1 | 0% | 1,175 | 1,213 | +3% | 0 | 0 | — |
case-08 | pass→pass | 7,098 | 2,803 | -61% | 1 | 1 | 0% | 1,120 | 1,173 | +5% | 0 | 0 | — |
case-23 | pass→pass | 19,036 | 21,226 | +12% | 1 | 1 | 0% | 2,605 | 3,788 | +45% | 0 | 0 | — |
case-24 | fail→pass | 15,822 | 18,505 | +17% | 1 | 1 | 0% | 2,218 | 3,392 | +53% | 0 | 0 | — |
case-01 | fail→fail | 26,043 | 56,380 | +116% | 1 | 1 | 0% | 3,737 | 4,612 | +23% | 0 | 0 | — |
case-02 | fail→fail | 26,137 | 24,917 | -5% | 1 | 1 | 0% | 3,638 | 4,510 | +24% | 0 | 0 | — |
case-03 | fail→pass | 26,301 | 24,711 | -6% | 1 | 1 | 0% | 3,911 | 4,530 | +16% | 0 | 0 | — |
case-04 | pass→pass | 23,018 | 24,846 | +8% | 1 | 1 | 0% | 3,230 | 4,471 | +38% | 0 | 0 | — |
case-05 | pass→pass | 16,247 | 18,170 | +12% | 1 | 1 | 0% | 2,458 | 3,440 | +40% | 0 | 0 | — |
case-06 | pass→pass | 18,809 | 17,881 | -5% | 1 | 1 | 0% | 2,557 | 3,413 | +33% | 0 | 0 | — |
case-09 | pass→pass | 21,225 | 19,514 | -8% | 1 | 1 | 0% | 2,948 | 3,738 | +27% | 0 | 0 | — |
case-10 | pass→pass | 22,747 | 20,836 | -8% | 1 | 1 | 0% | 3,286 | 3,775 | +15% | 0 | 0 | — |
case-11 | fail→pass | 20,472 | 20,727 | +1% | 1 | 1 | 0% | 3,003 | 3,981 | +33% | 0 | 0 | — |
case-12 | pass→pass | 18,219 | 17,272 | -5% | 1 | 1 | 0% | 2,458 | 3,249 | +32% | 0 | 0 | — |
case-13 | pass→pass | 23,391 | 20,703 | -11% | 1 | 1 | 0% | 3,332 | 3,819 | +15% | 0 | 0 | — |
case-14 | pass→pass | 16,340 | 19,496 | +19% | 1 | 1 | 0% | 2,247 | 3,533 | +57% | 0 | 0 | — |
case-15 | pass→pass | 17,838 | 22,331 | +25% | 1 | 1 | 0% | 2,638 | 4,058 | +54% | 0 | 0 | — |
case-16 | pass→pass | 14,752 | 20,673 | +40% | 1 | 1 | 0% | 2,360 | 3,807 | +61% | 0 | 0 | — |
case-17 | pass→pass | 18,752 | 19,008 | +1% | 1 | 1 | 0% | 2,817 | 3,713 | +32% | 0 | 0 | — |
case-18 | fail→pass | 9,821 | 17,903 | +82% | 1 | 1 | 0% | 1,359 | 3,473 | +156% | 0 | 0 | — |
case-19 | pass→pass | 19,272 | 16,230 | -16% | 1 | 1 | 0% | 2,844 | 3,052 | +7% | 0 | 0 | — |
case-20 | pass→pass | 16,825 | 15,507 | -8% | 1 | 1 | 0% | 2,486 | 3,102 | +25% | 0 | 0 | — |
case-21 | pass→pass | 32,262 | 18,550 | -43% | 1 | 1 | 0% | 2,309 | 3,270 | +42% | 0 | 0 | — |
case-22 | pass→pass | 21,021 | 27,418 | +30% | 1 | 1 | 0% | 3,180 | 4,797 | +51% | 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.