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，对客户工单进行分类归因、优先级排序与路由建议时使用本技能。适合输出分类分布、优先级清单与SLA风险提示。
.claude/skills/aifinlab-bank-t170-retail-finance-work-order-classification-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 58% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 68% | 0% |
本技能用于把零散的服务工单归类为可管理的类别与优先级，生成路由建议与复盘清单。强调分类口径一致、责任归口清晰、SLA风险可见。
time_window、work_ordersreferences/input-schema.mdbashpython scripts/run_skill.py --input assets/example-input.json --format markdown python scripts/run_skill.py --input assets/example-input.json --format json
脚本入口：scripts/run_skill.py（调用 shared/retail_service_ops_skill_engine.py 的 t170 场景）。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | fail→pass | 24,681 | 6,213 | -75% | 1 | 1 | 0% | 1,979 | 886 | -55% | 0 | 0 | — |
case-01 | fail→fail | 32,989 | 33,649 | +2% | 1 | 1 | 0% | 4,358 | 5,435 | +25% | 0 | 0 | — |
case-02 | fail→fail | 32,156 | 31,074 | -3% | 1 | 1 | 0% | 4,715 | 4,401 | -7% | 0 | 0 | — |
case-03 | fail→fail | 23,853 | 35,336 | +48% | 1 | 1 | 0% | 3,559 | 5,392 | +52% | 0 | 0 | — |
case-04 | fail→pass | 15,854 | 16,397 | +3% | 1 | 1 | 0% | 1,621 | 2,556 | +58% | 0 | 0 | — |
case-05 | fail→pass | 26,309 | 26,030 | -1% | 1 | 1 | 0% | 3,928 | 3,875 | -1% | 0 | 0 | — |
case-06 | fail→fail | 19,039 | 17,917 | -6% | 1 | 1 | 0% | 2,657 | 3,130 | +18% | 0 | 0 | — |
case-07 | pass→pass | 15,205 | 25,198 | +66% | 1 | 1 | 0% | 2,756 | 4,301 | +56% | 0 | 0 | — |
case-08 | fail→fail | 14,918 | 14,497 | -3% | 1 | 1 | 0% | 2,285 | 2,724 | +19% | 0 | 0 | — |
case-09 | fail→pass | 12,855 | 14,947 | +16% | 1 | 1 | 0% | 1,997 | 2,894 | +45% | 0 | 0 | — |
case-10 | pass→pass | 15,991 | 14,364 | -10% | 1 | 1 | 0% | 2,476 | 3,021 | +22% | 0 | 0 | — |
case-11 | pass→pass | 16,823 | 21,564 | +28% | 1 | 1 | 0% | 2,294 | 3,309 | +44% | 0 | 0 | — |
case-12 | pass→pass | 19,032 | 20,687 | +9% | 1 | 1 | 0% | 2,638 | 3,199 | +21% | 0 | 0 | — |
case-13 | fail→fail | 7,338 | 18,587 | +153% | 1 | 1 | 0% | 803 | 2,875 | +258% | 0 | 0 | — |
case-14 | pass→pass | 6,665 | 19,628 | +194% | 1 | 1 | 0% | 1,398 | 3,652 | +161% | 0 | 0 | — |
case-15 | pass→pass | 18,708 | 16,223 | -13% | 1 | 1 | 0% | 2,440 | 3,215 | +32% | 0 | 0 | — |
case-16 | fail→pass | 19,035 | 16,978 | -11% | 1 | 1 | 0% | 2,012 | 3,387 | +68% | 0 | 0 | — |
case-17 | fail→pass | 8,159 | 2,552 | -69% | 1 | 1 | 0% | 1,107 | 928 | -16% | 0 | 0 | — |
case-19 | fail→fail | 12,781 | 15,609 | +22% | 1 | 1 | 0% | 1,858 | 3,119 | +68% | 0 | 0 | — |
case-20 | fail→pass | 13,878 | 15,664 | +13% | 1 | 1 | 0% | 2,078 | 3,016 | +45% | 0 | 0 | — |
case-21 | pass→pass | 18,991 | 15,477 | -19% | 1 | 1 | 0% | 2,048 | 2,848 | +39% | 0 | 0 | — |
case-22 | fail→pass | 11,723 | 16,045 | +37% | 1 | 1 | 0% | 1,942 | 2,860 | +47% | 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 +36 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.