Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户在银行财富管理场景下需要做需求洞察、成因拆解、指标解读或专题诊断，并希望输出结构化结论与沟通提示时使用本技能。适合支持理财经理、私行顾问完成客户目标、风险偏好与市场环境的综合判断。
.claude/skills/aifinlab-bank-t194-wealth-management-needs-insight-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 56% | 0% |
本技能用于银行财富管理场景的“需求洞察”类分析，把客户目标、风险承受能力、期限与流动性约束、家庭资产结构、产品池与市场环境，转化为可沟通、可解释、可合规的诊断结论与后续动作建议。输出强调证据链、驱动项解释和后续追问方向，帮助理财经理/私行顾问推动下一步配置与沟通。
当需要把结构化输入快速生成标准化诊断草稿时，使用 scripts/needs_insight_builder.py：
脚本适合重复处理同类需求洞察材料，避免格式不一致或漏项。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | fail→fail | 22,666 | 24,932 | +10% | 1 | 1 | 0% | 2,792 | 4,406 | +58% | 0 | 0 | — |
case-01 | fail→fail | 30,903 | 34,961 | +13% | 1 | 1 | 0% | 3,857 | 4,697 | +22% | 0 | 0 | — |
case-02 | fail→fail | 25,293 | 27,187 | +7% | 1 | 1 | 0% | 3,224 | 4,386 | +36% | 0 | 0 | — |
case-03 | pass→pass | 24,827 | 15,676 | -37% | 1 | 1 | 0% | 3,916 | 2,998 | -23% | 0 | 0 | — |
case-04 | pass→pass | 24,530 | 26,937 | +10% | 1 | 1 | 0% | 3,611 | 4,218 | +17% | 0 | 0 | — |
case-05 | pass→pass | 15,101 | 30,784 | +104% | 1 | 1 | 0% | 2,190 | 4,398 | +101% | 0 | 0 | — |
case-06 | fail→pass | 25,759 | 34,179 | +33% | 1 | 1 | 0% | 3,286 | 5,286 | +61% | 0 | 0 | — |
case-07 | fail→fail | 23,866 | 23,102 | -3% | 1 | 1 | 0% | 2,825 | 4,057 | +44% | 0 | 0 | — |
case-08 | fail→pass | 21,522 | 25,523 | +19% | 1 | 1 | 0% | 3,100 | 4,308 | +39% | 0 | 0 | — |
case-09 | pass→pass | 23,036 | 28,720 | +25% | 1 | 1 | 0% | 2,907 | 4,577 | +57% | 0 | 0 | — |
case-10 | pass→pass | 21,667 | 20,176 | -7% | 1 | 1 | 0% | 2,893 | 3,862 | +33% | 0 | 0 | — |
case-11 | fail→fail | 24,364 | 22,031 | -10% | 1 | 1 | 0% | 3,079 | 4,080 | +33% | 0 | 0 | — |
case-12 | fail→fail | 22,918 | 28,109 | +23% | 1 | 1 | 0% | 2,891 | 4,481 | +55% | 0 | 0 | — |
case-13 | pass→pass | 25,432 | 23,661 | -7% | 1 | 1 | 0% | 3,216 | 3,905 | +21% | 0 | 0 | — |
case-14 | pass→pass | 23,219 | 23,625 | +2% | 1 | 1 | 0% | 3,119 | 4,167 | +34% | 0 | 0 | — |
case-15 | fail→pass | 21,588 | 30,389 | +41% | 1 | 1 | 0% | 3,016 | 4,559 | +51% | 0 | 0 | — |
case-16 | fail→fail | 23,289 | 27,583 | +18% | 1 | 1 | 0% | 3,018 | 4,712 | +56% | 0 | 0 | — |
case-17 | fail→pass | 21,232 | 25,010 | +18% | 1 | 1 | 0% | 3,066 | 4,085 | +33% | 0 | 0 | — |
case-18 | fail→pass | 25,034 | 30,291 | +21% | 1 | 1 | 0% | 2,909 | 4,547 | +56% | 0 | 0 | — |
case-20 | pass→pass | 16,163 | 23,837 | +47% | 1 | 1 | 0% | 2,133 | 3,820 | +79% | 0 | 0 | — |
case-21 | fail→pass | 22,717 | 62,190 | +174% | 1 | 1 | 0% | 3,061 | 3,819 | +25% | 0 | 0 | — |
case-22 | pass→pass | 25,028 | 25,966 | +4% | 1 | 1 | 0% | 3,164 | 4,467 | +41% | 0 | 0 | — |
case-23 | fail→fail | 31,593 | 25,737 | -19% | 1 | 1 | 0% | 3,137 | 4,296 | +37% | 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 +26 percentage points is the difference between those two pass rates over the 23 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.