/record-trade

description: Record a completed trade into all memory layers with full context
argument-hint: "[symbol] [direction] [entry] [exit] [pnl]"

Record Trade

Store a completed trade into TradeMemory with full context. This writes to all 5 OWM memory layers (episodic, semantic, procedural, affective, prospective) and updates behavioral baselines.

Workflow

Step 1: Gather Trade Details

If trade details are provided in the argument, parse them. Otherwise ask for:

• **Symbol**: e.g., XAUUSD, BTCUSDT, EURUSD
• **Direction**: long or short
• **Entry price** and **Exit price**
• **Lot size** (optional, defaults to strategy default)
• **P&L** in dollars

Step 2: Gather Context

Ask for or infer:

• **Strategy**: Which strategy triggered this trade (e.g., VolBreakout, IntradayMomentum)
• **Market context**: Session (London/NY/Asian), volatility regime, trend state
• **Reflection**: Why did you enter? What was the signal? Would you take it again?
• **Confidence**: 0.0-1.0, how confident were you at entry?

Step 3: Store via MCP

Use the `remember_trade` MCP tool to store across all OWM layers:

remember_trade({
  symbol, direction, entry_price, exit_price, pnl,
  strategy, lot_size, market_context, reflection, confidence
})

This automatically: 1. Creates an episodic memory (the raw trade event) 2. Updates semantic memory (strategy knowledge base) 3. Adjusts procedural memory (hold times, lot sizing patterns) 4. Updates affective state (confidence, drawdown, streak tracking) 5. Evaluates active prospective plans

Step 4: Confirm and Summarize

Report back:

• Trade stored successfully
• Updated affective state (new confidence level, streak)
• Any active trading plans that were affected
• Similar past trades (top 3 by OWM score) for quick comparison

Example

User: /record-trade XAUUSD long 5180 5210 +$150