Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在银行对公金融场景下，围绕客户画像进行成因拆解、指标解读、问题归因或专题诊断时使用本技能。适合输出结构化诊断结论、驱动项说明、后续追问方向和业务提示。
.claude/skills/aifinlab-bank-t145-corporate-finance-task-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 35% | 0% |
本技能面向银行对公金融场景，用于把“企业材料堆”压成一张可行动的客户画像：谁在控制、怎么赚钱、现金怎么回、为什么要融、风险可能在哪里、下一步该问什么/补什么/做什么。它不替代审批，但能显著提升尽调/授信/贷后/经营协同效率。
company.name + portrait.time_window + business.model_summaryreferences/input-schema.md如需把输入 JSON 自动渲染为“画像一页纸”，使用：
bashpython 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/corporate_ops_skill_engine.py 的 t145 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 | 27,093 | 35,782 | +32% | 1 | 1 | 0% | 3,214 | 5,611 | +75% | 0 | 0 | — |
case-02 | fail→pass | 21,653 | 28,286 | +31% | 1 | 1 | 0% | 3,068 | 4,668 | +52% | 0 | 0 | — |
case-03 | fail→fail | 29,652 | 30,649 | +3% | 1 | 1 | 0% | 3,800 | 5,212 | +37% | 0 | 0 | — |
case-04 | fail→fail | 27,436 | 28,650 | +4% | 1 | 1 | 0% | 3,817 | 5,019 | +31% | 0 | 0 | — |
case-05 | pass→pass | 24,658 | 32,869 | +33% | 1 | 1 | 0% | 3,536 | 5,421 | +53% | 0 | 0 | — |
case-06 | fail→pass | 29,693 | 40,654 | +37% | 1 | 1 | 0% | 3,659 | 5,453 | +49% | 0 | 0 | — |
case-07 | fail→pass | 20,054 | 27,608 | +38% | 1 | 1 | 0% | 2,790 | 4,214 | +51% | 0 | 0 | — |
case-08 | fail→pass | 21,696 | 26,466 | +22% | 1 | 1 | 0% | 3,288 | 4,417 | +34% | 0 | 0 | — |
case-09 | fail→fail | 21,455 | 29,351 | +37% | 1 | 1 | 0% | 3,139 | 5,485 | +75% | 0 | 0 | — |
case-10 | fail→pass | 26,276 | 27,618 | +5% | 1 | 1 | 0% | 3,260 | 4,388 | +35% | 0 | 0 | — |
case-11 | fail→pass | 25,697 | 35,417 | +38% | 1 | 1 | 0% | 3,368 | 5,342 | +59% | 0 | 0 | — |
case-12 | fail→pass | 24,126 | 31,811 | +32% | 1 | 1 | 0% | 3,303 | 4,852 | +47% | 0 | 0 | — |
case-13 | fail→fail | 21,960 | 26,718 | +22% | 1 | 1 | 0% | 3,224 | 4,827 | +50% | 0 | 0 | — |
case-14 | fail→fail | 22,616 | 27,177 | +20% | 1 | 1 | 0% | 3,249 | 4,401 | +35% | 0 | 0 | — |
case-15 | fail→fail | 24,446 | 25,864 | +6% | 1 | 1 | 0% | 3,203 | 4,934 | +54% | 0 | 0 | — |
case-20 | pass→pass | 15,064 | 19,495 | +29% | 1 | 1 | 0% | 2,379 | 3,420 | +44% | 0 | 0 | — |
case-21 | pass→pass | 27,502 | 28,326 | +3% | 1 | 1 | 0% | 3,372 | 4,542 | +35% | 0 | 0 | — |
case-16 | pass→fail | 24,181 | 27,725 | +15% | 1 | 1 | 0% | 2,925 | 5,087 | +74% | 0 | 0 | — |
case-17 | fail→fail | 21,900 | 29,569 | +35% | 1 | 1 | 0% | 3,253 | 4,580 | +41% | 0 | 0 | — |
case-18 | fail→fail | 24,421 | 34,038 | +39% | 1 | 1 | 0% | 2,886 | 5,341 | +85% | 0 | 0 | — |
case-19 | fail→pass | 33,680 | 28,353 | -16% | 1 | 1 | 0% | 3,149 | 5,175 | +64% | 0 | 0 | — |
case-22 | fail→pass | 7,425 | 13,047 | +76% | 1 | 1 | 0% | 454 | 2,570 | +466% | 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. 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.