TradingAgents Financial Framework Skill

This skill incorporates the TradingAgents multi-agent trading framework into the GenPark ecosystem. It provides the capability to simulate a complete trading firm using specialized LLM-powered agents.

Core Capabilities

The framework decomposes complex trading tasks into specialized roles, enabling a robust and scalable approach to market analysis and decision-making.

1. Analyst Team

Generates insights from multiple market perspectives.

• Fundamentals Analyst: Evaluates company financials, performance metrics, intrinsic values, and potential red flags.
• Sentiment Analyst: Analyzes social media and public sentiment using sentiment scoring to gauge short-term market mood.
• News Analyst: Monitors global news and macroeconomic indicators to interpret the impact of events.
• Technical Analyst: Utilizes technical indicators (MACD, RSI, etc.) to detect patterns and forecast price movements.

2. Researcher Team

Comprises bullish and bearish researchers who critically assess the Analyst Team's insights. Through structured debates, they balance potential gains against inherent risks to refine the analysis.

3. Trader Agent

Composes comprehensive reports from analysts and researchers to make informed trading decisions. It determines the optimal timing and magnitude of trades based on market insights.

4. Risk & Portfolio Management

• Risk Management: Continuously evaluates portfolio risk, assessing market volatility, liquidity, and strategy adjustments.
• Portfolio Manager: Reviews the risk assessment reports and approves or rejects transaction proposals. Approved orders are sent to a simulated exchange for execution.

System Architecture

• Framework Engine: Built on LangGraph to ensure flexibility and modularity.
• State Management:
• Decision Log: Persistently records trading decisions. On subsequent runs for a ticker, it fetches realized returns (raw and alpha vs. SPY), generates reflections, and injects past lessons into the Portfolio Manager's prompt.
• Checkpoint Resume: Utilizes SQLite to save state after each node, allowing crashed or interrupted runs to resume seamlessly from the last successful step.

LLM Provider Configuration

TradingAgents supports a wide array of LLMs. Set the appropriate API keys in the environment or configuration:

• Supported Providers: OpenAI (GPT), Google (Gemini), Anthropic (Claude), xAI (Grok), DeepSeek, Qwen (DashScope), GLM (Zhipu), OpenRouter.
• Local Models: Supported via Ollama.
• Enterprise Integrations: Azure OpenAI, AWS Bedrock.

Example Configuration Snippet

from tradingagents.graph.trading_graph import TradingAgentsGraph
from tradingagents.default_config import DEFAULT_CONFIG

config = DEFAULT_CONFIG.copy()
config["llm_provider"] = "openai"
config["deep_think_llm"] = "gpt-5.4" # Model for complex reasoning
config["quick_think_llm"] = "gpt-5.4-mini" # Model for quick tasks
config["max_debate_rounds"] = 2

ta = TradingAgentsGraph(debug=True, config=config)
_, decision = ta.propagate("NVDA", "2026-01-15")

Instructions for Use

1. Initialize the TradingAgents environment and install dependencies.
2. Select the target asset tickers and analysis dates.
3. Configure your preferred LLM provider for deep_think_llm and quick_think_llm.
4. Run the analysis and monitor the dynamic multi-agent debate.
5. Review the final Portfolio Manager decision and the generated decision logs for ongoing reflection.

[!WARNING] Trading performance may vary based on chosen backbone language models, model temperature, trading periods, data quality, and non-deterministic factors. This framework is designed for research purposes and is not intended as financial, investment, or trading advice.