Trade Mastrt Assistant Workflow
Overview
Trading Spark is a semi-automated trading assistant workflow running at second-level intervals, designed to capture and rigorously validate a trader's spontaneous trade ideas in real time. Each cycle, the workflow surfaces an interactive form prompting the trader to input a trading symbol, direction (Long / Short / Cover Long / Cover Short), quantity, and the rationale behind the idea. If no input is received within the allotted time, the workflow silently skips the current cycle and awaits the next trigger.
Core Logic
Once a trade idea is submitted, the workflow simultaneously launches three data collection branches:
- Market Sentiment Collection: Fetches real-time news and sentiment data for the target asset via an external data API. An AI large language model then performs both short-term and long-term sentiment scoring, producing quantified sentiment categories and confidence scores.
- Position Status Query: Automatically retrieves the account's current position direction and size in the target perpetual contract.
- Technical Indicator Calculation: Computes key technical indicators across four dimensions — trend (MACD), momentum (RSI), volatility (ATR), and volume (OBV).
Once all three data streams are merged, a second AI large language model conducts a comprehensive four-dimensional validation of the trade idea: rationale soundness, technical signal alignment, sentiment risk, and entry timing. The model outputs a complete 31-field trading decision report in strict flat JSON format, covering execution recommendation, confidence level, suggested entry price, stop-loss / take-profit targets, risk-reward ratio, and decision validity period.
Output & Logging
Each analysis result is appended to a persistent global trade log and exported as a CSV file saved locally, building a traceable history of all trade decisions.
Strategic Positioning
This workflow does not place orders autonomously. It acts as an "AI co-pilot" for the trader — completing data aggregation and multi-dimensional validation within seconds — helping traders identify cognitive biases, quantify risk, and refine entry logic, while keeping the human trader in full control of the final execution decision.
- 1
