Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when a wealth team needs to design high-net-worth client allocations, translate family goals and constraints into layered buckets, and output a compliant allocation plan with rebalancing steps and product shortlist.
.claude/skills/aifinlab-bank-t182-wealth-management-high-net-worth-allocation-high-net-worth-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 56% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 209% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 62% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 22% | 0% |
这个 skill 用于高净值客户的分层资产配置与方案落地，把家族目标、风险承受能力、流动性需求、币种结构与现有持仓统一成可执行、可复核、可沟通的配置建议。
scripts/run_skill.py：生成高净值分层配置建议与产品适配清单shared/wealth_management_skill_engine.py示例命令：
bashpython scripts/run_skill.py --input assets/example-input.json --format markdown
references/input-schema.mdreferences/output-schema.mdreferences/hnw-allocation-framework.mdreferences/liquidity-bucket-guidelines.mdreferences/risk-disclosure-points.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 36,597 | 43,658 | +19% | 1 | 1 | 0% | 4,838 | 6,505 | +34% | 0 | 0 | — |
case-02 | fail→fail | 42,651 | 39,550 | -7% | 1 | 1 | 0% | 5,537 | 6,182 | +12% | 0 | 0 | — |
case-03 | fail→pass | 34,798 | 42,233 | +21% | 1 | 1 | 0% | 4,657 | 7,248 | +56% | 0 | 0 | — |
case-04 | fail→fail | 31,780 | 32,823 | +3% | 1 | 1 | 0% | 4,269 | 5,607 | +31% | 0 | 0 | — |
case-05 | fail→pass | 36,753 | 32,514 | -12% | 1 | 1 | 0% | 5,239 | 5,446 | +4% | 0 | 0 | — |
case-06 | fail→pass | 8,924 | 20,267 | +127% | 1 | 1 | 0% | 1,099 | 3,398 | +209% | 0 | 0 | — |
case-07 | fail→pass | 26,327 | 29,977 | +14% | 1 | 1 | 0% | 3,497 | 5,666 | +62% | 0 | 0 | — |
case-08 | fail→fail | 19,020 | 28,379 | +49% | 1 | 1 | 0% | 3,154 | 5,306 | +68% | 0 | 0 | — |
case-09 | fail→fail | 26,411 | 33,089 | +25% | 1 | 1 | 0% | 3,670 | 5,360 | +46% | 0 | 0 | — |
case-19 | pass→pass | 31,450 | 36,223 | +15% | 1 | 1 | 0% | 3,908 | 5,628 | +44% | 0 | 0 | — |
case-10 | pass→pass | 25,468 | 25,938 | +2% | 1 | 1 | 0% | 3,470 | 4,139 | +19% | 0 | 0 | — |
case-11 | fail→pass | 27,043 | 26,052 | -4% | 1 | 1 | 0% | 3,841 | 4,671 | +22% | 0 | 0 | — |
case-12 | fail→pass | 24,144 | 33,962 | +41% | 1 | 1 | 0% | 3,673 | 6,260 | +70% | 0 | 0 | — |
case-13 | fail→pass | 27,999 | 38,750 | +38% | 1 | 1 | 0% | 3,715 | 6,045 | +63% | 0 | 0 | — |
case-14 | fail→pass | 28,327 | 28,680 | +1% | 1 | 1 | 0% | 3,683 | 5,227 | +42% | 0 | 0 | — |
case-15 | pass→pass | 17,965 | 22,883 | +27% | 1 | 1 | 0% | 2,517 | 3,877 | +54% | 0 | 0 | — |
case-16 | pass→pass | 26,151 | 32,249 | +23% | 1 | 1 | 0% | 3,667 | 6,044 | +65% | 0 | 0 | — |
case-17 | fail→pass | 24,001 | 37,177 | +55% | 1 | 1 | 0% | 3,878 | 6,366 | +64% | 0 | 0 | — |
case-18 | pass→pass | 26,489 | 30,719 | +16% | 1 | 1 | 0% | 3,628 | 5,352 | +48% | 0 | 0 | — |
case-20 | fail→pass | 30,542 | 39,257 | +29% | 1 | 1 | 0% | 3,980 | 6,350 | +60% | 0 | 0 | — |
case-21 | fail→pass | 26,984 | 30,943 | +15% | 1 | 1 | 0% | 3,232 | 4,874 | +51% | 0 | 0 | — |
case-22 | fail→fail | 30,092 | 53,559 | +78% | 1 | 1 | 0% | 3,791 | 7,760 | +105% | 0 | 0 | — |
case-23 | fail→pass | 25,104 | 27,381 | +9% | 1 | 1 | 0% | 3,202 | 4,905 | +53% | 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 +52 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.