Bridge between Binance trading events and TradeMemory Protocol. Automatically journals trades, recalls similar past setups, detects behavioral biases, and provides outcome-weighted recall for AI trading agents. Use this skill after executing Binance spot trades to build persistent memory.
Store Binance spot trades into persistent memory. Recall similar past trades before entering new positions. Detect behavioral biases (overtrading, revenge trading). Track strategy performance across sessions.
Requires: TradeMemory Protocol MCP server running.
Install and start the TradeMemory MCP server:
pip install tradememory-protocol
python -m tradememory
Or add to Claude Desktop / Claude Code MCP config:
{
"mcpServers": {
"tradememory": {
"command": "uvx",
"args": ["tradememory-protocol"]
}
}
}
After executing a Binance spot trade using the Binance Spot skill:
remember_trade MCP toolrecall_memories MCP toolget_agent_state to see if drawdown or confidence suggests pausingget_behavioral_analysis to detect biasesStore a completed trade into memory. Automatically updates all memory layers.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| symbol | string | Yes | Trading pair (e.g. "BTCUSDT", "ETHUSDT") |
| direction | string | Yes | "long" or "short" |
| entry_price | number | Yes | Entry price |
| exit_price | number | Yes | Exit price |
| pnl | number | Yes | Profit/loss in account currency |
| strategy_name | string | Yes | Strategy name (e.g. "GridBreakout", "MeanReversion") |
| market_context | string | Yes | Natural language description of market conditions |
| pnl_r | number | No | P&L as R-multiple (risk units) |
| context_regime | string | No | Market regime: trending_up, trending_down, ranging, volatile |
| confidence | number | No | Confidence level 0-1 (default 0.5) |
| reflection | string | No | Lessons learned from this trade |
Example — after a Binance spot BUY→SELL cycle:
Call remember_trade with:
symbol: "BTCUSDT"
direction: "long"
entry_price: 87500.00
exit_price: 89200.00
pnl: 170.00
strategy_name: "BreakoutEntry"
market_context: "BTC broke above 87000 resistance with volume spike. Funding rate positive. 4H RSI was 62."
context_regime: "trending_up"
confidence: 0.7
reflection: "Entry timing was good. Could have held longer — exited at first pullback."
Before entering a new trade, recall past trades in similar market conditions. Returns scored results ranked by outcome quality and context similarity.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| symbol | string | Yes | Trading pair to filter by |
| market_context | string | Yes | Current market conditions (natural language) |
| context_regime | string | No | Current regime: trending_up, trending_down, ranging, volatile |
| strategy_name | string | No | Filter by strategy |
| limit | number | No | Max results (default 10) |
Example — before entering a new BTC trade:
Call recall_memories with:
symbol: "BTCUSDT"
market_context: "BTC consolidating near 90000 after rally. Volume declining. Funding rate turning negative."
context_regime: "ranging"
strategy_name: "BreakoutEntry"
limit: 5
Returns past trades ranked by relevance to current conditions, with per-trade scores.
Check current trading state: confidence, risk appetite, drawdown, win/loss streaks.
No parameters required.
Returns a recommended action: normal, reduce_size, or stop_trading based on drawdown severity.
Detect trading biases from historical behavior patterns.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| strategy_name | string | No | Filter by strategy |
| symbol | string | No | Filter by symbol |
Detects: overtrading, revenge trading (re-entry after loss), disposition effect (cutting winners too early, holding losers too long), lot sizing inconsistency.
Get win rate, profit factor, and aggregate stats per strategy.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| strategy_name | string | No | Filter by strategy |
| symbol | string | No | Filter by symbol |
Set conditional plans that trigger on specific market conditions.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| trigger_type | string | Yes | "market_condition", "drawdown", or "time_based" |
| trigger_condition | string | Yes | JSON describing when to trigger |
| planned_action | string | Yes | JSON describing what to do |
| reasoning | string | Yes | Why this plan was created |
Example:
Call create_trading_plan with:
trigger_type: "market_condition"
trigger_condition: '{"regime": "volatile", "symbol": "BTCUSDT"}'
planned_action: '{"type": "reduce_size", "factor": 0.5}'
reasoning: "Historical data shows BreakoutEntry underperforms in volatile BTC regimes"
Check if any active plans match current market conditions.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| context_regime | string | No | Current market regime |
remember_trade with full context. Include market conditions, not just price data.recall_memories to check what happened in similar past conditions.get_agent_state to check if drawdown or losing streaks suggest reducing size.get_behavioral_analysis to detect emerging biases.market_context. The richer the description, the better the recall matching.TradeMemory Protocol is exchange-agnostic. While this skill documents the Binance bridge workflow, the same MCP tools work with any trading data source — just pass the correct symbol format for your exchange.
pnl_r (R-multiple) is optional but significantly improves recall quality.context_regime field enables regime-filtered recall — strongly recommended.