Multi-agent quantitative investment analysis system built with Microsoft Agent Framework (Semantic Kernel + AutoGen), using a Pregel-inspired data-flow workflow.
uv sync
cp .env.example .env # add Azure OpenAI credentials
uv run main.pyFetches stock data → generates technical signals (MACD, RSI) → backtests → reports metrics (CAGR, MDD, Sharpe Ratio).
Sample input
Analyze Apple (AAPL) stock using a momentum trading strategy:
1. Fetch historical data from 2023-01-01 to 2024-01-01
2. Generate buy/sell signals using MACD and RSI indicators
3. Backtest the strategy with initial capital of $10,000
4. Report performance metrics (CAGR, total return, final value)
Sample output
===== Final Output =====
Summary report — backtest outcome
Key results
- Final portfolio value: $11,157.97
- Total return: 11.58% (profit $1,157.97 on implied $10,000 start)
- CAGR: 11.68% (close to total return → consistent with ~1-year test horizon)
- Result files: backtest_results.xlsx / backtest_metrics.txt
Quick interpretation
- The strategy produced a positive return (~11.6%) on the test period with a final value of $11.16k.
- The near-equality of CAGR and total return indicates the backtest covers roughly one year.
- The absolute profit ($1,157.97) is modest but meaningful for a single-year horizon.
flowchart TD
classDef orchestrator fill:#f0fff0,stroke:#999,stroke-width:1px
classDef agent fill:#f0f8ff,stroke:#333,stroke-width:2px
classDef decision fill:#ffe4e1,stroke:#333,stroke-width:2px
A[User Input]:::orchestrator --> B[QuantInvestWorkflow]:::orchestrator
B --> C[WorkflowBuilder]:::orchestrator
subgraph Pipeline [Type-Safe Workflow Pipeline]
D[Stock Data Agent]:::agent --> E[Signal Generation Agent<br/>Python code executor]:::agent
E --> F{signals file?}:::decision
F -->|exists| G[Backtest Agent]:::agent
F -->|missing| H[Skip to Summary]:::agent
G --> I[Summary Report Agent]:::agent
H --> I
end
style Pipeline fill:none,stroke:#333,stroke-width:1px
C --> D
I --> J[Final Output]:::orchestrator
| Component | Description |
|---|---|
| Executors | Agents with tools — the workflow building blocks |
| Edges | Data-flow connections with conditional routing |
| WorkflowBuilder | Constructs the directed graph |
| Tools | agents/tools.py — function tools for each agent |
- All trading decisions are based on the previous day's signal (no look-ahead bias).
- Buy signal — daily return = change in Adjusted Close price.
- Sell signal — daily return =
(Open / Prev Day Close) − 1(captures overnight gap). - Consecutive identical signals are treated as Hold (no new trade opened).
- A sell cannot be executed without a prior buy.
| Rule | Behaviour |
|---|---|
| No prior buy | Sell signal skipped |
| Duplicate signal | Treated as Hold |
| Partial-year test | CAGR ≈ Total Return (expected) |
human_in_loop/main_invest_approval.py — human approval gate before executing investment decisions using RequestInfoExecutor.
flowchart TD
classDef human fill:#f0fff0,stroke:#999,stroke-width:1px
classDef agent fill:#f0f8ff,stroke:#333,stroke-width:2px
classDef decision fill:#ffe4e1,stroke:#333,stroke-width:2px
A[User Input<br/>Stock Ticker]:::human --> B[Investment Agent]:::agent
B --> C[Recommendation]:::agent
C --> D{Human Decision}:::decision
D -->|approve| E[Execute]:::agent
D -->|refine| F[Feedback]:::human
D -->|exit| G[Cancel]:::human
F --> B
uv run human_in_loop/main_invest_approval.py
MIT