Skip to main content

mcp_kit.generators.llm

LLM-based response generator for realistic mock responses.

LlmAuthenticationError​

class LlmAuthenticationError(Exception)

Exception raised when LLM authentication fails.

This exception is raised when the LLM service rejects the authentication credentials (API key, token, etc.).

LlmResponseGenerator​

class LlmResponseGenerator(ResponseGenerator)

Generate mock responses using an LLM agent.

This generator uses a Large Language Model to create realistic mock responses based on the tool context, making it suitable for testing scenarios that require believable synthetic data.

__init__​

def __init__(model: str)

Initialize the LLM response generator.

Arguments:

  • model: The LLM model identifier to use for generation

from_config​

@classmethod
def from_config(cls, config: DictConfig) -> Self

Create LLMResponseGenerator from configuration.

Arguments:

  • config: Configuration data with required 'model' parameter

Raises:

  • ValueError: If model parameter is missing from configuration

Returns:

LLMResponseGenerator instance

generate​

async def generate(target_name: str,
                   tool: Tool,
                   arguments: dict[str, Any] | None = None) -> list[Content]

Generate mock response using the LLM agent.

Creates a contextual prompt based on the tool information and uses the configured LLM to generate a realistic response.

Arguments:

  • target_name: Name of the target server
  • tool: The MCP tool definition
  • arguments: Arguments passed to the tool

Raises:

  • LlmAuthenticationError: If LLM authentication fails
  • ValueError: If the LLM response is empty

Returns:

List containing generated text content