Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要针对银行财富管理场景生成资产配置建议、客户配置方案、组合结构说明、 大类资产配置思路、再平衡建议、风险收益权衡分析时，使用本技能。 本技能适用于零售财富客户、私人银行客户、家族财富客户的配置建议初稿生成、 方案比较、持仓集中度分析、流动性安排建议和再平衡建议输出。 本技能强依赖行业数据与市场数据支持，包括但不限于宏观环境、利率水平、 估值区间、行业基准、同业配置趋势、波动率、相关性、回撤特征、产品分类口径等。 当行业数据或市场数据不足时，只能输出条件性建议、信息缺口说明和保守版配置框架， 不得将结果表述为确定性的投资建议。
.claude/skills/aifinlab-asset-allocation-advisory-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 68% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 98% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 94% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 92% | 0% |
本技能用于在银行财富管理场景下，围绕客户画像、风险承受能力、收益目标、期限偏好、 流动性需求、当前持仓结构与市场环境，生成资产配置建议、备选方案、再平衡建议与风险提示。
本技能不是投资承诺工具，不替代正式投资顾问审批、适当性管理、产品准入校验与合规披露流程。 当客户信息、市场数据、行业基准、同业对照或产品池信息不足时，应主动降级输出。
当出现以下任务时，应优先调用本技能：
以下情况不应直接输出明确配置比例结论：
优先收集以下信息：
输出时应至少包括：
识别客户风险等级、收益目标、期限偏好、流动性需求、家庭负担、持仓结构与集中度。 如客户核心资料缺失，应先列出待补充信息，不直接给出确定性比例。
重点识别：
综合以下数据进行校准：
若行业数据明显不足，只能输出区间型、情景型或原则型建议。
通常至少形成以下方案：
每套方案需说明：
给出阶段性调整建议：
评估资产是否足够覆盖短期刚性支出、紧急备用与已知负债需求。
判断当前与建议配置能否与客户收益目标相匹配，并解释目标是否过高。
判断可快速变现资产占比、锁定期限分布、赎回限制和现金流缺口。
分析资产类别、行业主题、期限、管理人、产品类型是否过度集中。
判断在当前利率、估值、波动环境下，配置建议是否具备合理性。
核验配置建议是否符合客户风险评级、准入范围与销售适当性要求。
本技能需要较强的行业数据支持，至少包括以下其中一部分：
若上述信息缺失，应在输出中明确写明：
优先按以下顺序输出：
references/asset_allocation_methodology.mdreferences/industry_data_requirements.mdreferences/suitability_and_compliance.mdreferences/output_schema.mdassets/templates/asset_allocation_report_template.mdassets/templates/information_gap_notice_template.mdassets/templates/customer_manager_talking_points_template.mdscripts/customer_profile_parser.pyscripts/allocation_recommendation_engine.pyscripts/render_asset_allocation_report.py| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 34,105 | 51,196 | +50% | 1 | 1 | 0% | 5,255 | 8,570 | +63% | 0 | 0 | — |
case-02 | fail→fail | 30,241 | 33,440 | +11% | 1 | 1 | 0% | 4,696 | 6,745 | +44% | 0 | 0 | — |
case-03 | fail→fail | 31,125 | 43,401 | +39% | 1 | 1 | 0% | 4,622 | 7,393 | +60% | 0 | 0 | — |
case-04 | pass→pass | 19,092 | 12,771 | -33% | 1 | 1 | 0% | 2,263 | 3,591 | +59% | 0 | 0 | — |
case-05 | pass→pass | 23,092 | 23,931 | +4% | 1 | 1 | 0% | 3,247 | 5,331 | +64% | 0 | 0 | — |
case-06 | pass→pass | 19,018 | 18,277 | -4% | 1 | 1 | 0% | 2,806 | 4,268 | +52% | 0 | 0 | — |
case-17 | pass→pass | 18,263 | 24,087 | +32% | 1 | 1 | 0% | 2,576 | 5,435 | +111% | 0 | 0 | — |
case-07 | fail→fail | 23,295 | 37,404 | +61% | 1 | 1 | 0% | 3,666 | 7,103 | +94% | 0 | 0 | — |
case-08 | fail→pass | 22,321 | 32,106 | +44% | 1 | 1 | 0% | 3,471 | 5,820 | +68% | 0 | 0 | — |
case-09 | fail→pass | 45,465 | 31,057 | -32% | 1 | 1 | 0% | 5,657 | 6,508 | +15% | 0 | 0 | — |
case-10 | pass→pass | 26,758 | 33,970 | +27% | 1 | 1 | 0% | 3,381 | 6,619 | +96% | 0 | 0 | — |
case-11 | pass→pass | 20,503 | 34,736 | +69% | 1 | 1 | 0% | 2,575 | 5,950 | +131% | 0 | 0 | — |
case-12 | fail→fail | 23,210 | 33,206 | +43% | 1 | 1 | 0% | 3,139 | 6,033 | +92% | 0 | 0 | — |
case-13 | fail→pass | 23,817 | 32,401 | +36% | 1 | 1 | 0% | 3,089 | 6,112 | +98% | 0 | 0 | — |
case-14 | pass→pass | 18,998 | 26,454 | +39% | 1 | 1 | 0% | 3,083 | 5,411 | +76% | 0 | 0 | — |
case-15 | fail→pass | 21,563 | 64,282 | +198% | 1 | 1 | 0% | 2,914 | 5,640 | +94% | 0 | 0 | — |
case-16 | pass→pass | 26,385 | 36,777 | +39% | 1 | 1 | 0% | 3,438 | 6,957 | +102% | 0 | 0 | — |
case-18 | pass→pass | 17,896 | 28,536 | +59% | 1 | 1 | 0% | 2,404 | 5,560 | +131% | 0 | 0 | — |
case-19 | pass→pass | 27,681 | 36,510 | +32% | 1 | 1 | 0% | 3,553 | 6,073 | +71% | 0 | 0 | — |
case-20 | pass→pass | 33,534 | 21,790 | -35% | 1 | 1 | 0% | 2,763 | 5,016 | +82% | 0 | 0 | — |
case-21 | pass→pass | 22,571 | 26,482 | +17% | 1 | 1 | 0% | 3,305 | 5,571 | +69% | 0 | 0 | — |
case-22 | fail→pass | 17,938 | 17,987 | +0% | 1 | 1 | 0% | 2,283 | 4,375 | +92% | 0 | 0 | — |
case-23 | pass→pass | 21,888 | 28,108 | +28% | 1 | 1 | 0% | 3,000 | 5,347 | +78% | 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 +22 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.