Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在交易银行与普惠金融场景下，对涉农客户准入进行初筛、资料核验与方案匹配时使用本技能。
.claude/skills/aifinlab-bank-t239-transaction-banking-inclusive-finance-admission-agri-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-23 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-25 | ✗→✓ | ▲ Improved | -34% | 0% |
本技能用于涉农客户准入初筛，强调农业经营周期、订单与补贴真实性、回款链路核验，并兼顾交易银行方案落地。
scripts/agri_admission_precheck.py：对涉农客户准入资料进行补贴与交易真实性校验，输出补件与风险提示。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | pass→pass | 17,286 | 20,878 | +21% | 1 | 1 | 0% | 2,575 | 3,329 | +29% | 0 | 0 | — |
case-01 | pass→pass | 46,671 | 52,413 | +12% | 1 | 1 | 0% | 2,526 | 3,305 | +31% | 0 | 0 | — |
case-02 | pass→pass | 17,169 | 16,480 | -4% | 1 | 1 | 0% | 2,396 | 2,770 | +16% | 0 | 0 | — |
case-03 | fail→pass | 22,036 | 51,012 | +131% | 1 | 1 | 0% | 3,148 | 3,418 | +9% | 0 | 0 | — |
case-04 | pass→pass | 13,598 | 17,252 | +27% | 1 | 1 | 0% | 2,166 | 3,166 | +46% | 0 | 0 | — |
case-05 | pass→pass | 17,004 | 19,167 | +13% | 1 | 1 | 0% | 2,528 | 3,297 | +30% | 0 | 0 | — |
case-06 | pass→pass | 25,031 | 18,929 | -24% | 1 | 1 | 0% | 3,538 | 3,059 | -14% | 0 | 0 | — |
case-07 | pass→pass | 49,243 | 18,719 | -62% | 1 | 1 | 0% | 2,810 | 3,194 | +14% | 0 | 0 | — |
case-09 | fail→pass | 23,937 | 20,640 | -14% | 1 | 1 | 0% | 3,701 | 3,571 | -4% | 0 | 0 | — |
case-10 | pass→pass | 19,114 | 18,103 | -5% | 1 | 1 | 0% | 2,608 | 3,110 | +19% | 0 | 0 | — |
case-11 | pass→pass | 17,494 | 18,353 | +5% | 1 | 1 | 0% | 2,514 | 3,175 | +26% | 0 | 0 | — |
case-12 | pass→pass | 22,308 | 22,508 | +1% | 1 | 1 | 0% | 3,133 | 3,729 | +19% | 0 | 0 | — |
case-13 | fail→fail | 21,761 | 20,524 | -6% | 1 | 1 | 0% | 3,042 | 3,452 | +13% | 0 | 0 | — |
case-14 | pass→pass | 20,921 | 19,796 | -5% | 1 | 1 | 0% | 3,075 | 3,210 | +4% | 0 | 0 | — |
case-15 | pass→pass | 21,316 | 22,456 | +5% | 1 | 1 | 0% | 3,059 | 3,532 | +15% | 0 | 0 | — |
case-16 | pass→pass | 19,072 | 17,669 | -7% | 1 | 1 | 0% | 2,600 | 2,979 | +15% | 0 | 0 | — |
case-17 | fail→pass | 22,150 | 18,306 | -17% | 1 | 1 | 0% | 3,156 | 3,110 | -1% | 0 | 0 | — |
case-18 | pass→pass | 19,422 | 17,853 | -8% | 1 | 1 | 0% | 2,960 | 2,991 | +1% | 0 | 0 | — |
case-19 | pass→pass | 20,390 | 19,101 | -6% | 1 | 1 | 0% | 2,810 | 3,295 | +17% | 0 | 0 | — |
case-20 | fail→fail | 22,809 | 19,578 | -14% | 1 | 1 | 0% | 3,000 | 3,193 | +6% | 0 | 0 | — |
case-21 | pass→pass | 16,255 | 17,813 | +10% | 1 | 1 | 0% | 2,365 | 2,990 | +26% | 0 | 0 | — |
case-22 | pass→pass | 40,417 | 21,056 | -48% | 1 | 1 | 0% | 2,976 | 3,491 | +17% | 0 | 0 | — |
case-23 | fail→pass | 20,470 | 14,742 | -28% | 1 | 1 | 0% | 2,911 | 2,479 | -15% | 0 | 0 | — |
case-24 | fail→fail | 25,028 | 23,263 | -7% | 1 | 1 | 0% | 3,788 | 3,830 | +1% | 0 | 0 | — |
case-25 | fail→pass | 15,816 | 7,861 | -50% | 1 | 1 | 0% | 2,497 | 1,637 | -34% | 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. 25 cases were attempted. The headline lift of +20 percentage points is the difference between those two pass rates over the 25 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.