SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsfund-deep-research

fund-deep-research

Data & analytics

View repository

Updated 10-15 19:55

About this skill

Fund In-Depth Due Diligence

Available Tools

This skill can call the following MCP data services, which can be selected as needed during the execution process:

Yingmi Financial Data (qieman)

  • Service address:https://dashscope.aliyuncs.com/api/v1/mcps/Qieman/sse
  • Core capabilities: fund search/diagnosis, portfolio analysis/backtesting, asset allocation plans, CFP toolchain, chart rendering
  • Main tools of this skill:SearchFunds, GuessFundCode, BatchGetFundsDetail, GetBatchFundPerformance, AnalyzeFundRisk, GetFundDiagnosis, BatchGetFundsHolding, getFundIndustryAllocation, getFundIndustryConcentration, getFundTurnoverRate, getFundBrinsonIndicator, getMarketTimingIndicator, getBondAllocationByFundCode, getBondFundCreditRatingLevel, getBondIndicator, getFundCampisiIndicator, RenderEchart

Hundsun Juyuan Financial Data (gildata-aidata)

  • Service address: obtained after activating the Hundsun Juyuan MCP service, in the format https://dashscope.aliyuncs.com/api/v1/mcps/<your-mcp-id>/mcp
  • Core capabilities: individual stock research (A/H/US), financial statements, capital flows, research reports and public opinion, wealth management products, macroeconomic data
  • Main tools of this skill:FundManagerInfoReport, FundManagerImageReport, ManagerProductsIncome, CompanyBasicInfo, ConsensusExpectation, FundAnnouncement

Core Principles

Charts first, concise text. Quantitative data such as performance comparison and industry allocation must be presented through RenderEchart generated visualization charts; text is only used to interpret key insights.

Input Requirements

Required Information

  • Fund code or fund name (provide at least one)

Optional Information

  • Analysis focus (such as "focus on risk" or "focus on manager capability")
  • Comparison benchmark or peer funds
  • Client risk level (affects suitability evaluation)

If the user provides a fund name but not a code, call GuessFundCode or SearchFunds to match.

Execution Process

Step 1: Determine the fund code and type

  • Via SearchFunds or GuessFundCode confirm the fund code
  • Call BatchGetFundsDetail to obtain basic fund information and determine the fund type

Step 2: Full-dimensional data collection (adjusted according to fund type)

General data (all types, qieman):

  • BatchGetFundsDetail: basic overview, scale, benchmark, risk level, manager information
  • GetBatchFundPerformance: returns across periods, performance analysis indicators
  • AnalyzeFundRisk: risk score, R-squared, residual variance
  • GetFundDiagnosis: comprehensive diagnosis, valuation, profit probability

Additional for equity/hybrid funds (qieman):

  • BatchGetFundsHolding: top ten heavy holdings
  • getFundIndustryAllocation: industry allocation distribution
  • getFundIndustryConcentration: industry concentration
  • getFundTurnoverRate: turnover rate
  • getFundBrinsonIndicator: Brinson return attribution (allocation/selection/interaction)
  • getMarketTimingIndicator: timing ability

Additional for bond funds (qieman):

  • getBondAllocationByFundCode: bond type allocation and style
  • getBondFundCreditRatingLevel: credit rating distribution
  • getBondIndicator: duration, leverage, concentration
  • getFundCampisiIndicator: Campisi return attribution

Step 3: In-depth fund manager profile (gildata-aidata)

  • FundManagerInfoReport: detailed information such as the manager's professional background, assets under management, and tenure
  • FundManagerImageReport: the manager's investment style profile (growth/value/balanced, large-cap/small-cap preference)
  • ManagerProductsIncome: a performance overview of all products managed by the manager, to evaluate their cross-product consistency

Step 4: Fundamental enhancement of top holdings (equity/hybrid, gildata-aidata)

For the top-ranked individual stocks among the top ten holdings (select the top 3-5), supplement:

  • CompanyBasicInfo: Company fundamental information (industry position, main business)
  • ConsensusExpectation: Market consensus expectations (earnings forecasts, target prices, rating distribution)

Step 5: Recent announcement search (gildata-aidata)

  • FundAnnouncement: Obtain the fund's recent important announcements (dividends, splits, manager changes, purchase restrictions, etc.)

Step 6: Four-dimensional diagnostic analysis

  • Performance dimension: Returns across each period, peer ranking, performance consistency
  • Risk dimension: Maximum drawdown, volatility, Sharpe/Calmar ratio, risk score
  • Holdings dimension(equity/hybrid): Concentration of top holdings, industry distribution, turnover rate, fundamental quality of top holdings
  • Manager capability dimension: Stock selection ability (Brinson), timing win rate, tenure, investment style profile, cross-product management capability

Step 7: Generate visualization charts (mandatory, cannot be skipped)

After data collection is complete,mustcall RenderEchart to generate the following charts:

Chart 1: Bar chart comparing performance across each period (must be generated)

Use the fund's and the peer average/benchmark's returns for each period to generate a grouped bar chart. ECharts option reference:

{
  "title": { "text": "业绩表现 vs 同类平均", "left": "center" },
  "tooltip": { "trigger": "axis" },
  "legend": { "top": "bottom" },
  "xAxis": { "type": "category", "data": ["近1月", "近3月", "近6月", "近1年", "近3年"] },
  "yAxis": { "type": "value", "axisLabel": { "formatter": "{value}%" } },
  "series": [
    { "name": "该基金", "type": "bar", "data": [2.1, 5.3, 8.7, 15.2, 42.1], "itemStyle": { "borderRadius": [4,4,0,0] } },
    { "name": "同类平均", "type": "bar", "data": [1.5, 3.8, 6.2, 10.1, 28.5], "itemStyle": { "borderRadius": [4,4,0,0] } }
  ]
}

Replace data with the actual return data.

Chart 2: Industry/bond type allocation pie chart (must be generated)

  • Equity/hybrid: Use getFundIndustryAllocation industry allocation data to generate a pie chart
  • Bond: Use getBondAllocationByFundCode bond type allocation data to generate a pie chart

ECharts option reference:

{
  "title": { "text": "行业配置分布", "left": "center" },
  "tooltip": { "trigger": "item", "formatter": "{b}: {d}%" },
  "legend": { "orient": "vertical", "left": "left", "top": "middle" },
  "series": [{
    "type": "pie",
    "radius": ["35%", "65%"],
    "center": ["58%", "50%"],
    "itemStyle": { "borderRadius": 6 },
    "label": { "formatter": "{b}\n{d}%" },
    "data": [
      { "name": "食品饮料", "value": 25.3 },
      { "name": "医药生物", "value": 18.1 },
      { "name": "电子", "value": 12.5 },
      { "name": "银行", "value": 10.2 },
      { "name": "其他", "value": 33.9 }
    ]
  }]
}

Replace data with the actual industry/bond type data. For bond funds, change the title to "Bond Type Allocation Distribution".

Chart 3: Four-dimensional risk-return radar chart (must be generated)

Use the four-dimensional diagnostic score data to generate a radar chart. ECharts option reference:

{
  "title": { "text": "四维诊断评分", "left": "center" },
  "radar": {
    "indicator": [
      { "name": "收益能力", "max": 100 },
      { "name": "风险控制", "max": 100 },
      { "name": "持仓质量", "max": 100 },
      { "name": "经理能力", "max": 100 }
    ],
    "shape": "circle"
  },
  "series": [{
    "type": "radar",
    "data": [{
      "value": [78, 65, 72, 80],
      "name": "综合评分",
      "areaStyle": { "opacity": 0.3 }
    }]
  }]
}

Based on data such as performance ranking, risk score, holdings concentration/industry diversification, and manager stock selection and timing ability, normalize each dimension to a 0-100 score and fill it in.

Step 8: Suitability assessment

  • If the user has provided the client's risk level, assess whether the fund is suitable
  • Provide a conclusion on "what type of investors it is suitable for"

Output template

Output according to the following structure,embed charts in the corresponding sections, and keep each text section to 2-4 sentences：

## 基金深度尽调 | {基金名称}({基金代码})

### 一、基本信息
| 项目 | 详情 |
|------|------|
| 基金类型 | {类型} |
| 成立日期 | {日期} |
| 最新规模 | {X}亿 |
| 基金经理 | {姓名}（任职{X}年） |
| 风险等级 | {R等级} |

### 二、四维诊断

{四维诊断雷达图}

### 三、业绩表现

{业绩对比柱状图}

{2-3句解读：各阶段表现特征、同类排名水平}

### 四、持仓分析

{行业/券种配置饼图}

{2-3句解读：集中度、行业偏好、风格特征}

**重仓股基本面速览**（股票型/混合型）：
| 重仓股 | 行业 | 市场一致预期 |
|--------|------|------------|
| {股票名} | {行业} | {盈利预测/评级} |

### 五、基金经理画像
- **投资风格**：{成长/价值/均衡}，{大盘/小盘偏好}
- **跨产品表现**：管理{X}只产品，整体业绩{评价}
- **选股能力**：Brinson 归因 {评价}
- **择时胜率**：{X}%

### 六、风险评估
- **最大回撤**：{X}%（{评价}）
- **夏普比率**：{X}（{评价}）
- **风险评分**：{X}/100

### 七、近期动态
{近期重要公告摘要，如无重大公告可省略}

### 八、综合结论
**尽调评级：{推荐/中性/谨慎}**
- 优势：{1-2点}
- 风险点：{1-2点}
- 适合：{适配客户类型}

---
*尽调报告基于历史数据分析，不构成投资建议。基金有风险，投资需谨慎。*

Notes

  • Charts are mandatory: The performance comparison bar chart, allocation pie chart, and four-dimensional radar chart are required items
  • Compliance bottom line: Direct investment advice wording such as "recommend buying" or "suggest increasing position" must not appear
  • Objective and neutral: Both advantages and risk points must be mentioned; do not beautify one-sidedly
  • Type adaptation: Automatically adjust the analysis framework according to the fund type (equity/bond/hybrid/QDII)
  • Concise text: Keep the full text to 800-1200 words (excluding charts), with each section not exceeding 4 sentences
  • Data source division of labor: Use qieman for fund dimension data, and gildata-aidata for manager profiles and fundamentals of top holdings