Grader

What grader MCP server does

The grader MCP server is described as a tool for evaluating trading strategies through backtesting. Its output is an A-F grade intended to communicate how a strategy performs under the stated evaluation approach. The available description specifically emphasizes real out-of-sample data, which distinguishes the service from an assessment based solely on the data used to develop a strategy.

The intended use is straightforward: provide or assess a trading strategy, run it against out-of-sample market data, and use the resulting letter grade as a compact evaluation signal. The available material does not explain the strategy input format, supported asset classes, supported markets, time periods, or the precise grading criteria. Those details should be confirmed in the linked API documentation before implementation.

The grader MCP server may fit workflows where an agent needs to compare or screen trading strategies using a stated backtesting result. It can also serve as a review step after strategy development, provided the strategy and data requirements match the API’s undocumented details. The listing does not establish that the service executes trades, supplies investment advice, connects to a brokerage, or manages a portfolio.

Limitations and notes

The repository README is unavailable in the provided material. As a result, there is no grounded information here about installation commands, MCP transport, authentication, environment variables, request schemas, response fields, available tool names, rate limits, pricing, or client-specific setup. The linked source is the Tessen API documentation, but no documentation content was supplied beyond that URL and the short listing description.

The grader MCP server should therefore be treated as an evaluation service rather than a complete trading system. The confirmed capability is strategy grading through backtests on real out-of-sample data. Nothing in the available material confirms live execution, paper trading, broker integration, market-data access beyond the backtesting claim, or automated strategy generation.

Before adopting it, verify how strategies are submitted, what “out-of-sample” means for the service, which markets and instruments are supported, and how the A-F result is calculated. Also confirm whether the API exposes an MCP-native interface or requires an adapter. No named MCP clients are supported by the supplied information, so client compatibility cannot be confirmed from this listing.