Smart Fund Investment Advisor and Portfolio Configuration Assistant

Specific Problem: Common Pain Points for Individual Fund Investors

Individual investors often grapple with several core issues in fund investing: lacking professional asset allocation skills to build portfolios aligned with their risk tolerance and goals; having fragmented holdings information that makes real-time tracking of NAV changes and return attribution difficult; finding it inefficient to manually screen fund managers or products amid complex market information; and lacking systematic tools for scenario simulation or strategy backtesting during market volatility. Traditional tools are often single-function or lack interactivity, failing to seamlessly connect analysis, allocation, and decision-support workflows.

How the Skill Works: Core Capabilities and Key Steps

This is a locally-run Python package simulating the workflow of an AI client manager. Its core capabilities include:

• Portfolio Construction and Optimization: Employs a three-tier SAA/TAA (Strategic/Tactical Asset Allocation) framework, with asset_allocator.py implementing models like risk parity and mean-variance. Users can get configuration suggestions, including core-satellite portfolios, rebalancing plans, and fee estimates, based on questionnaires or direct risk input. The v10.2 portfolio optimizer (portfolio_enhancer.py) can process return matrices to compute the efficient frontier.
• Quantitative Backtesting and Multi-Factor Analysis: Provides backtest_engine.py to run historical backtests with strategies like momentum or mean reversion. factor_engine.py analyzes portfolio exposures across six dimensions (momentum, volatility, value, etc.) and compares with peers, aiding in diagnosing style drift.
• Interaction and Dialogue Capabilities: Supports two main interaction methods. First, integration with external AI assistants via the MCP protocol (52 tools) for natural language queries, such as calling query_fund or compare_managers. Second, the v10.2 zero-dependency voice interaction (voice_interface.py), launched via python -m fund_advisor voice, accepts spoken commands to query holdings or goal progress. manager_dialogue.py can also distill public views from fund managers to simulate conversational responses.
• Post-Investment Monitoring and Report Generation: performance_tracker.py provides a six-dimensional dashboard for tracking, while review_engine.py generates weekly/monthly review reports. Users can customize reports including performance, attribution, and risk modules using the modular report generator (report_generator.py), exporting to PPT/PDF. goal_tracker.py quantitatively tracks the probability of achieving investment goals.

Key steps typically involve: 1) Importing holdings via holdings_importer.py (supports screenshots, Excel, etc.); 2) Running asset allocation analysis or portfolio optimization; 3) Setting up periodic reviews or using the voice assistant for queries; 4) Iterating strategies by backfilling results from advice logged in the learning/ module.

Application Boundaries and Important Notes

This skill is positioned as a personal learning and analytical reference tool. First, it does not constitute any investment advice; all analysis is based on historical data and models, with no guarantee of future performance. Second, its capabilities are strictly limited to the fund domain and are unsuitable for short-term stock trading or non-fund financial management. When using quantitative analysis and scenario simulation, users must understand the model assumptions and limitations. The voice interaction and manager dialogue functions rely on local rules or an optional LLM, and their distilled views do not represent real-time opinions from the fund managers themselves. All data reconstruction and updates can be completed with zero external dependencies, but runtime considerations apply.