MCP server for options trading strategy analysis and execution via TastyTrade API. Provides market overview scanning, strategy building (12 strategy types), position management, and trade execution with OAuth authentication.
TastyScanner has well-structured tool definitions with clear naming (verb_noun pattern), comprehensive parameter schemas using Zod, and detailed descriptions. However, output schemas are not documented, error handling lacks recovery guidance, and some parameter descriptions could be more specific about constraints and formats. The server demonstrates solid fundamentals but falls short of production-grade polish in response documentation and error messaging.
Adjust an open order's limit price. Supports 'improve_fill' (auto-adjust by one tick) or 'custom' (specify custom_price).
Close an open position by position ID. Supports take_profit, stop_loss, dte_expiry, or manual close reasons.
Execute a multi-leg options trade on TastyTrade. Requires explicit legs array with OCC symbols, limit price, and price effect (Credit/Debit). Only executable if ENABLE_LIVE_TRADING=true.
Fetch account balance, buying power, and cash position from TastyTrade.
Check TastyTrade connection status, account number, and market status.
Scan symbols for options trading metrics: price, IVR (0-100, higher=better for selling premium), IV, beta, earnings date. Sorted by IVR descending. Pass watchlist name, explicit symbols array, or omit for defaults (~40 popular tickers). Returns max 50 results.
Output schemas not documented. Tool descriptions state what is returned (e.g., 'Returns max 50 results', 'Returns position details') but do not specify the structure of response objects, field types, or nested properties. LLMs cannot plan downstream tool calls or extract specific fields without knowing the response schema.
Error handling lacks recovery guidance. Error responses (via makeError() helper) return code and message but do not tell the LLM what to do next or whether the error is retryable. E.g., 'Connection failed' does not indicate if the agent should retry, check credentials, or ask the user.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 67 | 2026-07-28+ | v2 |
Fetch all open positions in the TastyTrade account. Returns position details including strategy type detection, legs, P&L, and days to expiration.
Build options strategies for a symbol. Supports 12 strategy types: iron_condor, put_credit_spread, call_credit_spread, iron_butterfly, jade_lizard, twisted_sister, long_straddle, long_strangle, bull_call_spread, bear_put_spread, call_butterfly, put_butterfly. Pass 'all' to generate all types. Returns sorted by risk/reward ratio.
Fetch all personal and platform watchlists available in the TastyTrade account. Returns watchlist names and entry counts.
Destructive operations (execute_trade, close_position) lack confirmation or dry-run support. An agent can execute a live trade or close a position without a confirmation step, risking unintended financial consequences. No pattern for user confirmation before irreversible actions.
Parameter descriptions lack specific constraints. E.g., min_ivr is described as '(0-100)' in the tool description but the parameter description omits the range. limit_price for execute_trade has no guidance on format (cents vs dollars, decimal places). Agents cannot validate inputs without explicit constraints in parameter descriptions.
Getter tools (get_positions, get_watchlists, get_account_balance) have minimal descriptions (70 chars or less). Descriptions do not explain when to call them, what data they reveal, or how they fit into a trading workflow. LLMs cannot distinguish between similar discovery tools without richer context.