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，围绕运营活动进行客户分群、节奏编排、渠道触达和转化跟踪时使用本技能。适合输出运营名单、活动排期、A/B策略和监控指标。
.claude/skills/aifinlab-bank-t152-retail-finance-customer-operation-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 84% | 0% |
| case-23 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 190% | 0% |
本技能用于银行零售客户经营的结构化分析与动作生成，把客户清单转成“可执行名单 + 可追踪动作 + 可复盘指标”。强调活动排期、渠道选择、频控和转化漏斗复盘。
eferences/input-schema.md
eferences/output-schema.md
ash python 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_customer_ops_skill_engine.py 的 $(System.Collections.Hashtable.id) 场景。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 29,038 | 86,507 | +198% | 1 | 1 | 0% | 4,381 | 4,976 | +14% | 0 | 0 | — |
case-02 | fail→fail | 29,771 | 24,708 | -17% | 1 | 1 | 0% | 4,001 | 4,592 | +15% | 0 | 0 | — |
case-03 | fail→pass | 16,118 | 16,310 | +1% | 1 | 1 | 0% | 2,581 | 2,897 | +12% | 0 | 0 | — |
case-04 | pass→pass | 17,088 | 18,506 | +8% | 1 | 1 | 0% | 2,718 | 3,546 | +30% | 0 | 0 | — |
case-05 | pass→pass | 11,611 | 25,597 | +120% | 1 | 1 | 0% | 1,591 | 4,608 | +190% | 0 | 0 | — |
case-06 | pass→pass | 17,128 | 23,296 | +36% | 1 | 1 | 0% | 2,510 | 3,745 | +49% | 0 | 0 | — |
case-07 | pass→pass | 19,296 | 18,708 | -3% | 1 | 1 | 0% | 2,321 | 3,609 | +55% | 0 | 0 | — |
case-08 | pass→pass | 20,626 | 29,713 | +44% | 1 | 1 | 0% | 2,624 | 4,466 | +70% | 0 | 0 | — |
case-09 | pass→pass | 16,955 | 23,670 | +40% | 1 | 1 | 0% | 2,419 | 3,803 | +57% | 0 | 0 | — |
case-10 | pass→pass | 27,233 | 19,373 | -29% | 1 | 1 | 0% | 3,563 | 3,652 | +2% | 0 | 0 | — |
case-11 | pass→pass | 22,350 | 29,028 | +30% | 1 | 1 | 0% | 2,896 | 4,500 | +55% | 0 | 0 | — |
case-12 | pass→pass | 21,902 | 21,533 | -2% | 1 | 1 | 0% | 2,678 | 4,039 | +51% | 0 | 0 | — |
case-13 | pass→pass | 17,288 | 21,220 | +23% | 1 | 1 | 0% | 2,432 | 3,764 | +55% | 0 | 0 | — |
case-14 | fail→fail | 18,890 | 19,617 | +4% | 1 | 1 | 0% | 2,821 | 3,645 | +29% | 0 | 0 | — |
case-15 | fail→pass | 15,168 | 16,931 | +12% | 1 | 1 | 0% | 1,924 | 3,541 | +84% | 0 | 0 | — |
case-16 | pass→pass | 20,573 | 19,456 | -5% | 1 | 1 | 0% | 2,680 | 3,273 | +22% | 0 | 0 | — |
case-17 | pass→pass | 18,063 | 20,932 | +16% | 1 | 1 | 0% | 2,284 | 3,784 | +66% | 0 | 0 | — |
case-18 | pass→pass | 20,889 | 19,236 | -8% | 1 | 1 | 0% | 2,629 | 3,228 | +23% | 0 | 0 | — |
case-19 | pass→pass | 21,292 | 23,916 | +12% | 1 | 1 | 0% | 3,362 | 3,806 | +13% | 0 | 0 | — |
case-20 | pass→pass | 19,058 | 18,595 | -2% | 1 | 1 | 0% | 3,240 | 3,526 | +9% | 0 | 0 | — |
case-21 | pass→pass | 15,808 | 20,604 | +30% | 1 | 1 | 0% | 2,592 | 3,316 | +28% | 0 | 0 | — |
case-22 | pass→pass | 32,216 | 26,578 | -18% | 1 | 1 | 0% | 4,204 | 4,722 | +12% | 0 | 0 | — |
case-23 | fail→pass | 21,416 | 19,406 | -9% | 1 | 1 | 0% | 3,134 | 3,688 | +18% | 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. 23 cases were attempted. The headline lift of +13 percentage points is the difference between those two pass rates over the 23 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.