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，围绕组合推荐进行适配、配置、推荐、方案设计或备选方案比较时使用本技能。适合输出主方案、备选方案、匹配逻辑、风险提示和沟通要点。
.claude/skills/aifinlab-bank-t147-corporate-finance-task-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 64% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 282% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 128% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 55% | 0% |
本技能用于在对公场景下做“可落地”的产品组合推荐：先把需求与约束说清楚，再把产品从“清单”变成“方案”（交易结构、材料、节点、风险提示、备选切换条件）。输出强调可解释、可审批、可执行。
company.name + needs.goal + needs.constraintsreferences/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 t147 场景）。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→fail | 20,578 | 25,313 | +23% | 1 | 1 | 0% | 2,616 | 3,979 | +52% | 0 | 0 | — |
case-01 | fail→fail | 24,974 | 30,556 | +22% | 1 | 1 | 0% | 3,753 | 5,181 | +38% | 0 | 0 | — |
case-02 | fail→fail | 33,299 | 27,123 | -19% | 1 | 1 | 0% | 4,310 | 4,878 | +13% | 0 | 0 | — |
case-03 | fail→pass | 19,471 | 19,241 | -1% | 1 | 1 | 0% | 2,122 | 3,483 | +64% | 0 | 0 | — |
case-04 | pass→pass | 6,488 | 31,303 | +382% | 1 | 1 | 0% | 924 | 2,891 | +213% | 0 | 0 | — |
case-05 | pass→pass | 19,398 | 21,882 | +13% | 1 | 1 | 0% | 2,269 | 3,442 | +52% | 0 | 0 | — |
case-06 | fail→pass | 28,062 | 32,209 | +15% | 1 | 1 | 0% | 3,740 | 4,851 | +30% | 0 | 0 | — |
case-07 | fail→pass | 7,288 | 25,733 | +253% | 1 | 1 | 0% | 1,220 | 4,664 | +282% | 0 | 0 | — |
case-08 | fail→pass | 13,502 | 29,690 | +120% | 1 | 1 | 0% | 1,851 | 4,212 | +128% | 0 | 0 | — |
case-09 | fail→fail | 21,550 | 30,002 | +39% | 1 | 1 | 0% | 2,778 | 4,374 | +57% | 0 | 0 | — |
case-11 | fail→pass | 22,641 | 28,845 | +27% | 1 | 1 | 0% | 2,804 | 4,342 | +55% | 0 | 0 | — |
case-12 | pass→pass | 22,790 | 37,900 | +66% | 1 | 1 | 0% | 3,013 | 5,804 | +93% | 0 | 0 | — |
case-13 | pass→pass | 19,341 | 30,744 | +59% | 1 | 1 | 0% | 2,892 | 5,389 | +86% | 0 | 0 | — |
case-14 | fail→fail | 22,017 | 27,455 | +25% | 1 | 1 | 0% | 3,586 | 4,953 | +38% | 0 | 0 | — |
case-15 | pass→pass | 21,540 | 28,301 | +31% | 1 | 1 | 0% | 2,761 | 4,328 | +57% | 0 | 0 | — |
case-16 | pass→fail | 23,818 | 31,461 | +32% | 1 | 1 | 0% | 3,210 | 5,551 | +73% | 0 | 0 | — |
case-17 | fail→fail | 45,418 | 26,914 | -41% | 1 | 1 | 0% | 3,550 | 4,644 | +31% | 0 | 0 | — |
case-18 | fail→pass | 22,957 | 23,277 | +1% | 1 | 1 | 0% | 2,788 | 3,896 | +40% | 0 | 0 | — |
case-19 | pass→pass | 23,559 | 29,209 | +24% | 1 | 1 | 0% | 3,032 | 4,565 | +51% | 0 | 0 | — |
case-20 | fail→pass | 23,857 | 42,713 | +79% | 1 | 1 | 0% | 3,712 | 4,955 | +33% | 0 | 0 | — |
case-21 | fail→pass | 24,050 | 25,207 | +5% | 1 | 1 | 0% | 3,233 | 4,300 | +33% | 0 | 0 | — |
case-22 | pass→pass | 19,208 | 30,880 | +61% | 1 | 1 | 0% | 2,727 | 5,156 | +89% | 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 +32 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.