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，围绕机会识别进行持续监测、风险扫描、异常识别或预警提示时使用本技能。适合输出风险信号摘要、优先级判断、处置建议和升级路径。
.claude/skills/aifinlab-bank-t146-corporate-finance-task-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 86% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 28% | 0% |
本技能用于把监测信号（交易、税票、工商司法、舆情、内部规则命中等）快速翻译成两件事：机会（该做什么业务动作）与风险（该怎么排查、何时升级）。核心不是堆“告警列表”，而是输出可执行的跟进路径。
company.name + monitoring.time_window + monitoring.signalsreferences/input-schema.mdbashpython scripts/run_skill.py --input assets/example-input.json --format markdown
脚本入口：scripts/run_skill.py（调用 shared/corporate_ops_skill_engine.py 的 t146 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→pass | 22,075 | 28,778 | +30% | 1 | 1 | 0% | 3,223 | 5,155 | +60% | 0 | 0 | — |
case-02 | fail→fail | 26,246 | 29,172 | +11% | 1 | 1 | 0% | 3,439 | 4,410 | +28% | 0 | 0 | — |
case-03 | pass→pass | 15,358 | 15,935 | +4% | 1 | 1 | 0% | 2,158 | 3,006 | +39% | 0 | 0 | — |
case-04 | fail→pass | 15,527 | 22,483 | +45% | 1 | 1 | 0% | 2,126 | 3,954 | +86% | 0 | 0 | — |
case-05 | pass→pass | 21,444 | 23,234 | +8% | 1 | 1 | 0% | 2,527 | 3,549 | +40% | 0 | 0 | — |
case-06 | fail→pass | 21,044 | 21,631 | +3% | 1 | 1 | 0% | 2,617 | 3,754 | +43% | 0 | 0 | — |
case-07 | fail→pass | 29,377 | 38,977 | +33% | 1 | 1 | 0% | 3,502 | 5,139 | +47% | 0 | 0 | — |
case-08 | fail→fail | 26,556 | 24,897 | -6% | 1 | 1 | 0% | 3,317 | 3,882 | +17% | 0 | 0 | — |
case-09 | pass→pass | 23,617 | 21,807 | -8% | 1 | 1 | 0% | 2,795 | 3,765 | +35% | 0 | 0 | — |
case-10 | fail→fail | 23,315 | 20,334 | -13% | 1 | 1 | 0% | 2,994 | 3,909 | +31% | 0 | 0 | — |
case-11 | fail→pass | 24,554 | 27,731 | +13% | 1 | 1 | 0% | 3,205 | 4,103 | +28% | 0 | 0 | — |
case-12 | fail→pass | 23,982 | 24,214 | +1% | 1 | 1 | 0% | 3,491 | 4,495 | +29% | 0 | 0 | — |
case-13 | pass→pass | 29,551 | 20,613 | -30% | 1 | 1 | 0% | 3,377 | 3,726 | +10% | 0 | 0 | — |
case-14 | fail→pass | 17,683 | 23,707 | +34% | 1 | 1 | 0% | 2,198 | 3,605 | +64% | 0 | 0 | — |
case-15 | fail→fail | 23,768 | 22,846 | -4% | 1 | 1 | 0% | 3,339 | 4,073 | +22% | 0 | 0 | — |
case-16 | fail→pass | 24,448 | 21,395 | -12% | 1 | 1 | 0% | 3,108 | 3,844 | +24% | 0 | 0 | — |
case-17 | fail→fail | 22,951 | 27,823 | +21% | 1 | 1 | 0% | 3,562 | 4,186 | +18% | 0 | 0 | — |
case-18 | fail→fail | 26,121 | 30,647 | +17% | 1 | 1 | 0% | 3,597 | 4,683 | +30% | 0 | 0 | — |
case-19 | fail→pass | 24,596 | 25,135 | +2% | 1 | 1 | 0% | 3,042 | 3,944 | +30% | 0 | 0 | — |
case-20 | fail→pass | 30,471 | 33,763 | +11% | 1 | 1 | 0% | 3,594 | 5,314 | +48% | 0 | 0 | — |
case-21 | fail→fail | 27,072 | 25,457 | -6% | 1 | 1 | 0% | 3,172 | 4,160 | +31% | 0 | 0 | — |
case-22 | pass→pass | 24,270 | 23,768 | -2% | 1 | 1 | 0% | 3,047 | 4,009 | +32% | 0 | 0 | — |
case-23 | fail→pass | 20,692 | 22,537 | +9% | 1 | 1 | 0% | 2,936 | 3,676 | +25% | 0 | 0 | — |
case-24 | fail→pass | 25,896 | 31,062 | +20% | 1 | 1 | 0% | 3,073 | 4,684 | +52% | 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. 24 cases were attempted. The headline lift of +50 percentage points is the difference between those two pass rates over the 24 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.