Loading ToolBench

Reading the current corpus…

mcp-market-data-server MCP server · C · ToolBench

mcp-market-data-server

github.com/rtavarezz/mcp-market-data-server

A FastAPI-based MCP server that provides market data tools for stock prices, cryptocurrency prices, and stock statistics, backed by a PostgreSQL database for price ingestion and historical queries.

<=2025-11-25STREAMABLE_HTTP3 tools

61/100

Definition Quality84

Protocol Readiness62

Supportability21

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

Three tools with clear naming (verb_noun pattern: get_stock_price, get_crypto_price, get_stock_stats) and well-structured Pydantic schemas. Descriptions are concise and actionable (43-94 chars, within the 10-1024 range). All parameters have type definitions and descriptions. Input schemas are fully documented via Pydantic models. Output schemas explicitly defined with typed response classes. However, parameter descriptions are minimal (single phrases without usage guidance), and no error guidance for common failure modes (invalid ticker, API failures, rate limits). Missing per-parameter constraints (e.g., ticker format, symbol enums) and no documentation of error recovery paths. Respects READ_ONLY risk classification.

Tools (3)

get_crypto_priceread onlysource verified84/100

Return the current USD price for a cryptocurrency symbol.

get_stock_priceread onlysource verified84/100

Return the latest price for a stock ticker.

get_stock_statsread onlysource verified83/100

Return today's open/high/low/close and percentage change for a stock ticker.

Issues (4)

MEDIUMtool-description ↗

Parameter descriptions are minimal and lack usage context. 'Stock ticker symbol, e.g., AAPL' is present but does not explain format constraints, valid ranges, or when to use this tool vs alternatives.

get_stock_priceget_crypto_priceget_stock_stats

HIGHconstrained-input ↗

No input validation constraints documented. Parameters accept free-form strings without specifying format (e.g., ticker must be 1-5 uppercase letters, symbol must match known exchanges). LLMs may pass invalid values like 'AAPL1' or 'btc-usd' without guidance.

get_stock_priceget_crypto_priceget_stock_stats

HIGHrecovery-guide ↗

Error responses are defined (SymbolNotFoundError, RateLimitError, UpstreamServiceError, MarketDataError observed in code) but the HTTP error responses do not include recovery guidance. 'Symbol not found' should suggest 'Try a different ticker' or list available symbols.

get_stock_priceget_crypto_priceget_stock_stats

Recommendations

Add format constraints to parameter descriptions: 'ticker: Stock ticker symbol (1-5 uppercase letters, e.g., AAPL, MSFT). Do not include exchange prefix.'
Enrich error messages with recovery steps. Change SymbolNotFoundError response from 'Symbol not found' to 'Symbol INVALID not found. Common symbols: AAPL, MSFT, GOOGL, AMZN. Call list_symbols() to discover available tickers.'
Document expected data freshness and latency in tool descriptions. E.g., 'Returns latest price (updated every 15 seconds during market hours).'
Add a 'source' field to output schemas (e.g., 'source: "real_time" | "delayed_15min" | "cached"') so agents know data freshness.
Consider adding optional parameters for time-based queries: add an optional 'as_of' parameter to get_stock_stats to retrieve historical data, or document whether tools only return current/today's data.
Add idempotent hints to tool definitions (via tool annotations) since these are read-only operations with no side effects.

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

↑ 61 points across a rubric change (v1 → v2)

61/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-22	C	61	<=2025-11-25	v2
2026-03-09	F	0	-	v1