mutiagent4financial

Generate alpha signals using momentum agent.

Discover alpha factors.

Develop strategy configuration.

Run comprehensive backtest.

Submit strategy to memory.

Run integrated backtest pipeline.

Validate strategy performance.

Get the status of a specific agent or all agents. Args: agent_id: Optional agent ID. If not provided, returns all agents. Returns: Dictionary containing agent status information.

Start a new agent in the pool. Args: agent_id: Unique identifier for the agent agent_type: Type of agent to start (e.g., 'momentum', 'trend_following') config: Optional configuration for the agent Returns: Dictionary containing the result of the start operation.

Stop a running agent. Args: agent_id: ID of the agent to stop graceful: Whether to perform a graceful shutdown Returns: Dictionary containing the result of the stop operation.

Restart an existing agent. Args: agent_id: ID of the agent to restart Returns: Dictionary containing the result of the restart operation.

Get comprehensive health information for the entire pool. Returns: Dictionary containing pool health metrics and status.

Get detailed performance metrics for the pool. Returns: Dictionary containing performance analytics.

Update configuration for a running agent. Args: agent_id: ID of the agent to configure config: New configuration parameters Returns: Dictionary containing the result of the configuration update.

Stores a structured memory in the Neo4j graph database with intelligent linking and semantic indexing.

Stores multiple memories in a batch operation for high-throughput processing with enhanced performance monitoring.

Retrieves memories using enhanced full-text search with intelligent ranking and semantic capabilities.

Filters memories based on structured criteria with enhanced query capabilities and performance optimization.

Retrieves comprehensive statistics with enhanced metrics including performance analytics and system health.

Retrieves memories with intelligent relationship expansion for comprehensive context discovery.

Performs intelligent semantic search using AI embeddings, similarity scoring, and context understanding.

Extracts and analyzes trending keywords from recent memories using intelligent text processing.

Intelligently deletes old and irrelevant memories with enhanced criteria and safety checks.

Publishes memory-related events to the real-time stream processing system for reactive analytics.

Returns comprehensive real-time statistics from the stream processing system and event analytics.

Fetch and store market data in shared context.

Store a learned lesson into Neo4j long-term memory for future reference.

Query past lessons from Neo4j memory. Use keywords like 'overfitting', 'momentum', 'risk'.

Search the Alpha101 research paper (RAG) for relevant factor construction knowledge.

Execute the complete standard alpha pipeline (Alpha158 Factors -> Train Model -> Generate Signals). Uses 200+ Alpha158 factors by default with LightGBM model.

Calculate specific technical indicators on the current data (Test Data).

Train a model using TRAINING data and predict on CURRENT features. Requires 'train_data' in context.

Convert predictions to trading signals and finalize the task.

Get comprehensive gateway status and health information

List all internal agents and their connection status

Orchestrate a task across multiple internal agents

Get the status and results of an orchestrated task

Make direct call to internal agent via MCP

Generate comprehensive alpha strategy using internal agents

Coordinate memory synchronization across agents

Emergency shutdown of gateway and internal agents

Store a learned lesson into the Neo4j Long-term memory graph.

Query past lessons from Neo4j Long-term memory using a keyword like 'overfitting'.

Retrieve relevant alpha factor research from the Alpha101 paper using RAG/VectorDB.

Load and analyze asset data, compute technical indicators and signals

Compute and evaluate IC/IR using Alpha158 handler

Propose alpha factors based on Qlib results

Generate iteration report

List all registered sub-agents.

Check the health of the momentum agent.

Get a value by key via A2A memory if supported.

Set a key-value via A2A memory if supported.

Delete a key via A2A memory if supported.

List keys via A2A memory if supported.

Submit an AlphaTaskDTO to the Orchestrator

Cancel a pending or running task

List all currently active tasks

Get orchestrator performance metrics

Run RL backtest and update agent policy for a given symbol and market data.

Process strategy requests and generate alpha signals

Submit strategy flow events to memory system for tracking and analysis.

Generate comprehensive academic-style strategy research report

Store comprehensive analysis results including factors, strategy, and backtest data

Get A2A memory connection status

Execute standard portfolio construction pipeline.

Construct portfolio weights from signals in context.

Submit final portfolio.

Execute the full multi-dimensional risk pipeline and store results in context.

Calculate per-stock and portfolio volatility. Stores in ctx.risk_metrics.

Generate LLM risk narrative from the metrics already in context.

Submit final risk assessment from metrics accumulated in context.

Get the current account summary including buying power, cash, and portfolio value.

Get all current open positions. Returns a JSON list.

Execute a list of orders with risk validation.

Cancel all pending orders on Alpaca.

Get recent order history from Alpaca.

Receive input from external orchestrator and autonomously create tasks

Query memory agent for relevant knowledge

Dynamically generate code tools based on analysis requirements

Create validation programs for generated code

Execute previously generated code tools

Generate trading signal and strategy flow output

Fetch historical price and volume data from external data sources. Supports multiple data providers: - polygon: Polygon.io API (high-frequency, institutional-grade) - yfinance: Yahoo Finance (free, daily data) Returns OHLCV data aligned to the requested frequency as JSON string. Automatically handle…

Compute an alpha factor using a specified formula on price/volume data. Supports: - Mean reversion factors (z-score, residual-based) - Technical indicators (RSI, Bollinger Bands) - Custom formulas using pandas/numpy syntax Returns a time series of factor values with proper alignment and handling of …

Backtest an alpha factor with realistic transaction costs and holding periods. Evaluates: - Information Coefficient (IC): Rank correlation between signals and forward returns - Sharpe Ratio: Risk-adjusted return metric - Turnover: Portfolio rebalancing frequency - Transaction costs: Slippage and com…

Visualize backtest results and factor performance. Generates: - Cumulative returns chart - IC time series - Factor exposure distribution - Performance attribution Returns plot URLs or base64-encoded images. Args: metrics: JSON string with performance metrics output_format: Output format for visualiz…

Generate a sophisticated alpha strategy flow using LLM analysis and real market data.

Analyze market sentiment using LLM for the given symbol.

Store a memory record with intelligent linking and semantic indexing

Retrieve memories using enhanced search capabilities

Perform AI-powered semantic search across memories

Get comprehensive system statistics and health information

Perform comprehensive health check of all system components

Create intelligent relationships between memory nodes