Loading ToolBench
Finnhub market data MCP server for quotes, fundamentals, technical analysis, news sentiment, and crypto workflows.
15 tools with significant definition gaps. Most tools have descriptions but lack parameter-level documentation. Schemas are present but incomplete: many parameters lack type definitions or descriptions. Tool naming follows verb_noun convention (good), but parameter schemas are underdeveloped. No output schemas documented. Error handling absent. Security concerns: no indication of secret injection for API keys. Composition is problematic: tools like stock_market_data bundle 7 operations into one tool, violating single-responsibility principle.
Alternative data operations including ESG scores, social sentiment, supply chain data, and patents.
Calendar data operations including IPO calendar, earnings calendar, economic calendar, and FDA calendar.
Cryptocurrency data operations including exchanges, symbols, quotes, and candles.
Forex data operations including exchanges, symbols, rates, and candles.
Get job status.
Market events operations including market holidays, analyst ratings, and merger/acquisition data.
Multi-operation tools violate single-responsibility principle. stock_market_data bundles 7 operations (get_quote, get_candles, get_company_profile, symbol_lookup, get_basic_financials, get_financials_as_reported, get_earnings_surprises) into one tool. This forces LLMs to reason about which operation to invoke and makes error handling ambiguous.
Parameters lack type definitions and descriptions. Example: stock_market_data has 'resolution', 'from', 'to', 'query', 'metricType', 'freq' parameters with minimal or no descriptions. 'from' and 'to' lack type info (number vs string). 'resolution' has no description of valid values (e.g., '1', '5', '15', '60', 'D', 'W', 'M').
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 44 | <=2025-11-25 | v2 |
News and sentiment operations including company news, market news, news sentiment, and insider sentiment.
Create a new project.
List all projects.
SEC filings operations including SEC filings retrieval, filing sentiment analysis, and similarity index.
Stock estimates operations including earnings estimates, revenue estimates, EBITDA estimates, price targets, and recommendation trends.
Stock fundamentals operations including basic financials, financials as reported, dividends, stock splits, and revenue breakdown.
Stock market data operations including quotes, candles, company profiles, symbol lookup, basic financials, financials as reported, and earnings surprises.
Stock ownership operations including insider transactions, institutional ownership, institutional portfolio, and congress transactions.
Technical analysis operations including technical indicators, aggregate signals, pattern recognition, and support/resistance levels.
No output schemas documented. Tools return data from Finnhub API but no schema describes what fields agents should expect. LLMs cannot plan downstream calls or extract required data without knowing response structure.
No error handling guidance. Tools provide no recovery hints. If a symbol lookup fails, agents have no guidance on what to try next (e.g., 'Try search_symbols() with a partial name'). Raw API errors are likely returned without actionable context.
API key exposure risk. No evidence of secret injection pattern. Finnhub API key likely passed as environment variable but no documentation of secure handling. If key appears in tool parameters or responses, it will leak into agent traces and logs.