MCP Tools Reference

SkopaqTrader exposes 23 MCP tools through skopaq/mcp_server.py. All tools are async and return JSON strings.

Market Data

Tool	Description	Key Args
`get_quote`	Real-time LTP, OHLC, bid/ask, volume, change%	`symbol`
`get_historical`	OHLCV candles (last 20 returned)	`symbol`, `days=5`, `resolution=1`

Example — get a quote:

What is the current price of RELIANCE?

Claude calls mcp__skopaq__get_quote(symbol="RELIANCE") and returns:

{
  "symbol": "RELIANCE",
  "ltp": 2485.50,
  "open": 2470.00,
  "high": 2498.00,
  "low": 2465.00,
  "change_pct": 0.63
}

Example — historical data:

Show me 15-minute candles for TCS over the last 3 days.

mcp__skopaq__get_historical(symbol="TCS", days=3, resolution=15)

Portfolio Management

Tool	Description	Key Args
`get_positions`	Open positions with P&L	none
`get_holdings`	Delivery (CNC) holdings	none
`get_funds`	Available cash, margin, collateral	none
`get_orders`	Today's orders with status	none

These tools try Kite Connect first. If Kite is not connected, they fall back to the paper engine.

AI Analysis

Tool	Description	Key Args
`analyze_stock`	Full 15-agent pipeline (2-5 min)	`symbol`, `date=""`
`scan_market`	Multi-model market scan	`max_candidates=5`
`check_safety`	Pre-trade safety validation	`symbol`, `quantity`, `price`, `side`

Analysis duration

analyze_stock runs the complete multi-agent pipeline with 4 analysts, bull/bear debate, risk debate, and trader decision. It takes 2-5 minutes and calls multiple LLM providers.

Data Gathering (Claude-Native)

These tools fetch raw data for Claude to reason over directly, bypassing the multi-LLM pipeline:

Tool	Description	Key Args
`gather_all_analysis_data`	One-shot: market + news + fundamentals + social + memories	`symbol`, `date=""`
`gather_market_data`	OHLCV + RSI, MACD, Bollinger, SMA, EMA, ATR, VWMA	`symbol`, `date=""`
`gather_news_data`	Company news, global macro, insider transactions	`symbol`, `date=""`
`gather_fundamentals_data`	Profile, balance sheet, cash flow, income statement	`symbol`, `date=""`
`gather_social_data`	Social media sentiment and company news	`symbol`, `date=""`
`recall_agent_memories`	BM25 search over past trade lessons	`situation_summary`
`quick_decision`	Calibrated answer to a question about a text via TypeSafe Jev (~0.1 s; needs `SKOPAQ_JEV_ENABLED`)	`text`, `question`, `options=None`
`save_trade_reflection`	Store post-trade lesson for future reference	`symbol`, `side`, `entry_price`, `exit_price`, `pnl`, `pnl_pct`

Order Execution

Tool	Description	Key Args
`place_order`	Execute order through safety checker	`symbol`, `side="BUY"`, `quantity=1`, `price=0`, `order_type="MARKET"`
`system_status`	Version, mode, token health, active LLMs	none

Safety First

Every order passes through the SafetyChecker before execution. Orders that violate position limits, daily loss caps, or other safety rules are rejected automatically.

Options Trading

Tool	Description	Key Args
`get_option_chain`	Full chain with calls/puts, OI, volume, distance%	`symbol="NIFTY"`, `expiry_index=0`
`suggest_option_trade`	AI strike selection with risk metrics	`symbol="NIFTY"`, `strategy="SHORT_PUT"`, `expiry_index=0`

Supported strategies:

SHORT_PUT -- Sell OTM put (bullish view)
SHORT_CALL -- Sell OTM call (bearish view)
SHORT_STRANGLE -- Sell OTM put + call (neutral view)

GTT Orders

Tool	Description	Key Args
`place_gtt_order`	GTT buy trigger or OCO sell (target + stop-loss)	`symbol`, `action`, `trigger_price`, `target_price`, `stop_loss_price`, `quantity`
`list_gtt_orders`	List all active GTT orders	none
`setup_swing_trade`	Complete CNC swing: GTT buy + planned OCO sell	`symbol`, `entry_price`, `target_price`, `stop_loss_price`, `quantity`

GTT orders require Kite Connect

GTT (Good Till Triggered) orders are a Zerodha feature. You must be logged into Kite via /api/kite/login before using these tools.

Tool Count by Category

Category	Count
Market Data	2
Portfolio	4
AI Analysis	3
Data Gathering	7
Execution	2
Options	2
GTT / Swing	3
Total	23