Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在银行交易银行与普惠场景下，围绕结算方案进行适配、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-t244-transaction-banking-inclusive-finance-settlement-solution-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 89% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -37% | 0% |
本技能面向交易银行与普惠金融场景，既要兼顾准入效率和材料真实性，也要把结算、现金管理、供应链和贸易融资方案写到能落地的层面。 当前这支 skill 更偏向服务 交易银行和普惠团队，输出时要特别注意 兼顾准入效率、方案落地和风险边界。
本技能提供结算方案生成脚本，基于业务场景与账户体系输出可执行结算路径。
scripts/settlement_solution_builder.py用途：根据交易链路与账户体系生成结算方案与备选方案。
输入字段（JSON）：
business_flow: 交易链路与对手方settlement_needs: 结算频次、币种、清算方式account_structure: 账户体系与额度限制constraints: 合规与系统对接约束输出内容：
命令行示例：
python scripts/settlement_solution_builder.py --input assets/settlement_input.json --output outputs/settlement_solution.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 59,760 | 54,826 | -8% | 1 | 1 | 0% | 4,296 | 4,776 | +11% | 0 | 0 | — |
case-02 | fail→pass | 27,990 | 64,044 | +129% | 1 | 1 | 0% | 3,822 | 5,904 | +54% | 0 | 0 | — |
case-03 | fail→fail | 28,958 | 27,890 | -4% | 1 | 1 | 0% | 4,370 | 5,188 | +19% | 0 | 0 | — |
case-04 | pass→pass | 6,715 | 13,588 | +102% | 1 | 1 | 0% | 893 | 2,846 | +219% | 0 | 0 | — |
case-05 | pass→pass | 10,387 | 20,250 | +95% | 1 | 1 | 0% | 1,551 | 3,718 | +140% | 0 | 0 | — |
case-06 | fail→pass | 17,150 | 25,184 | +47% | 1 | 1 | 0% | 2,608 | 4,922 | +89% | 0 | 0 | — |
case-07 | fail→pass | 6,920 | 3,295 | -52% | 1 | 1 | 0% | 1,279 | 1,619 | +27% | 0 | 0 | — |
case-08 | fail→pass | 16,544 | 34,464 | +108% | 1 | 1 | 0% | 2,958 | 1,857 | -37% | 0 | 0 | — |
case-09 | fail→pass | 19,503 | 32,046 | +64% | 1 | 1 | 0% | 2,684 | 4,382 | +63% | 0 | 0 | — |
case-10 | fail→fail | 17,632 | 23,075 | +31% | 1 | 1 | 0% | 2,482 | 4,164 | +68% | 0 | 0 | — |
case-11 | fail→fail | 13,443 | 23,935 | +78% | 1 | 1 | 0% | 1,937 | 4,375 | +126% | 0 | 0 | — |
case-12 | pass→pass | 19,632 | 25,432 | +30% | 1 | 1 | 0% | 2,724 | 4,397 | +61% | 0 | 0 | — |
case-13 | fail→fail | 13,129 | 23,121 | +76% | 1 | 1 | 0% | 2,001 | 4,479 | +124% | 0 | 0 | — |
case-14 | fail→fail | 21,235 | 25,825 | +22% | 1 | 1 | 0% | 2,963 | 4,738 | +60% | 0 | 0 | — |
case-15 | fail→pass | 18,417 | 18,759 | +2% | 1 | 1 | 0% | 3,182 | 3,767 | +18% | 0 | 0 | — |
case-16 | pass→pass | 21,128 | 31,363 | +48% | 1 | 1 | 0% | 2,968 | 5,086 | +71% | 0 | 0 | — |
case-17 | fail→fail | 17,733 | 21,580 | +22% | 1 | 1 | 0% | 2,504 | 4,118 | +64% | 0 | 0 | — |
case-18 | fail→pass | 22,096 | 23,410 | +6% | 1 | 1 | 0% | 3,167 | 4,500 | +42% | 0 | 0 | — |
case-19 | fail→fail | 20,461 | 19,134 | -6% | 1 | 1 | 0% | 2,831 | 3,798 | +34% | 0 | 0 | — |
case-20 | fail→pass | 18,823 | 25,550 | +36% | 1 | 1 | 0% | 2,691 | 4,869 | +81% | 0 | 0 | — |
case-21 | fail→pass | 21,624 | 27,817 | +29% | 1 | 1 | 0% | 3,222 | 5,205 | +62% | 0 | 0 | — |
case-22 | fail→pass | 17,920 | 20,150 | +12% | 1 | 1 | 0% | 2,348 | 3,918 | +67% | 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 +50 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.