Skip to main content
Glama
prompts.py12.9 kB
"""Prompt testing tools for MCP server testing. This module provides MCP tools for discovering and retrieving prompts from connected target MCP servers, enabling comprehensive prompt testing workflows. """ import logging import time from typing import Annotated, Any from fastmcp import Context from ..connection import ConnectionError, ConnectionManager from ..mcp_instance import mcp logger = logging.getLogger(__name__) @mcp.tool async def list_prompts(ctx: Context) -> dict[str, Any]: """List all prompts available on the connected MCP server. Retrieves comprehensive information about all prompts exposed by the target server, including names, descriptions, and complete argument schemas to enable accurate prompt invocation. Returns: Dictionary with prompt listing including: - success: True on successful retrieval - prompts: List of prompt objects with name, description, and arguments schema - metadata: Total count, server info, timing information Raises: Returns error dict if not connected or retrieval fails """ start_time = time.perf_counter() try: # Verify connection exists client, state = ConnectionManager.require_connection() # User-facing progress update await ctx.info("Listing prompts from connected MCP server") # Detailed technical log logger.info("Listing prompts from connected MCP server") # Get prompts from the server prompts_result = await client.list_prompts() elapsed_ms = (time.perf_counter() - start_time) * 1000 # Convert prompts to dictionary format with full argument schemas # Note: client.list_prompts() returns a list directly, not an object with .prompts prompts_list = [] for prompt in prompts_result: # Extract arguments schema arguments = [] if hasattr(prompt, "arguments") and prompt.arguments: for arg in prompt.arguments: arg_dict = { "name": arg.name, "description": arg.description if arg.description else "", "required": arg.required if hasattr(arg, "required") else False, } arguments.append(arg_dict) prompt_dict = { "name": prompt.name, "description": prompt.description if prompt.description else "", "arguments": arguments, } prompts_list.append(prompt_dict) metadata = { "total_prompts": len(prompts_list), "server_url": state.server_url, "retrieved_at": time.time(), "request_time_ms": round(elapsed_ms, 2), } # Add server info if available if state.server_info: metadata["server_name"] = state.server_info.get("name", "unknown") metadata["server_version"] = state.server_info.get("version") # User-facing success update await ctx.info(f"Retrieved {len(prompts_list)} prompts from server") # Detailed technical log logger.info( f"Retrieved {len(prompts_list)} prompts from server", extra={ "prompt_count": len(prompts_list), "server_url": state.server_url, "duration_ms": elapsed_ms, }, ) return { "success": True, "prompts": prompts_list, "metadata": metadata, } except ConnectionError as e: elapsed_ms = (time.perf_counter() - start_time) * 1000 # User-facing error update await ctx.error(f"Not connected: {str(e)}") # Detailed technical log logger.error(f"Not connected: {str(e)}", extra={"duration_ms": elapsed_ms}) return { "success": False, "error": { "error_type": "not_connected", "message": str(e), "details": {}, "suggestion": "Use connect_to_server() to establish a connection first", }, "prompts": [], "metadata": { "request_time_ms": round(elapsed_ms, 2), }, } except Exception as e: elapsed_ms = (time.perf_counter() - start_time) * 1000 # User-facing error update await ctx.error(f"Failed to list prompts: {str(e)}") # Detailed technical log logger.exception("Failed to list prompts", extra={"duration_ms": elapsed_ms}) # Increment error counter ConnectionManager.increment_stat("errors") return { "success": False, "error": { "error_type": "execution_error", "message": f"Failed to list prompts: {str(e)}", "details": {"exception_type": type(e).__name__}, "suggestion": "Check that the server supports the prompts capability and is responding correctly", }, "prompts": [], "metadata": { "request_time_ms": round(elapsed_ms, 2), }, } @mcp.tool async def get_prompt( name: Annotated[str, "Name of the prompt to retrieve"], arguments: Annotated[dict[str, Any], "Dictionary of arguments to pass to the prompt"], ctx: Context ) -> dict[str, Any]: """Get a rendered prompt from the connected MCP server. Retrieves a prompt by name with the provided arguments and returns the rendered prompt messages. Returns: Dictionary with rendered prompt including: - success: True if prompt was retrieved successfully - prompt: Object with name, description, and rendered messages - metadata: Request timing and server information Raises: Returns error dict for various failure scenarios: - not_connected: No active connection - prompt_not_found: Prompt doesn't exist on server - invalid_arguments: Arguments don't match prompt schema - execution_error: Prompt retrieval failed """ start_time = time.perf_counter() try: # Verify connection exists client, state = ConnectionManager.require_connection() # User-facing progress update await ctx.info(f"Getting prompt '{name}' with arguments") # Detailed technical log logger.info( f"Getting prompt '{name}' with arguments", extra={"prompt_name": name, "arguments": arguments}, ) # Get the prompt prompt_start = time.perf_counter() result = await client.get_prompt(name, arguments) prompt_elapsed_ms = (time.perf_counter() - prompt_start) * 1000 # Increment statistics ConnectionManager.increment_stat("prompts_executed") total_elapsed_ms = (time.perf_counter() - start_time) * 1000 # Extract prompt messages messages = [] if hasattr(result, "messages") and result.messages: for message in result.messages: message_dict: dict[str, Any] = { "role": message.role, } # Handle different content types # Content can be: TextContent, ImageContent, AudioContent, ResourceLink, EmbeddedResource if hasattr(message, "content"): content = message.content if hasattr(content, "type"): # Structured content with type discriminator content_dict: dict[str, Any] = { "type": content.type, } # Handle type-specific fields if content.type == "text" and hasattr(content, "text"): content_dict["text"] = content.text elif content.type == "image" and hasattr(content, "data"): content_dict["data"] = content.data if hasattr(content, "mimeType"): content_dict["mimeType"] = content.mimeType elif content.type == "audio" and hasattr(content, "data"): content_dict["data"] = content.data if hasattr(content, "mimeType"): content_dict["mimeType"] = content.mimeType elif content.type == "resource": # ResourceLink or EmbeddedResource if hasattr(content, "uri"): content_dict["uri"] = content.uri if hasattr(content, "resource"): content_dict["resource"] = content.resource message_dict["content"] = content_dict else: # Fallback for simple/unknown content types message_dict["content"] = {"type": "text", "text": str(content)} messages.append(message_dict) prompt_info = { "name": name, "description": result.description if hasattr(result, "description") and result.description else "", "messages": messages, } # User-facing success update await ctx.info(f"Prompt '{name}' retrieved successfully with {len(messages)} messages") # Detailed technical log logger.info( f"Prompt '{name}' retrieved successfully", extra={ "prompt_name": name, "message_count": len(messages), "duration_ms": prompt_elapsed_ms, }, ) return { "success": True, "prompt": prompt_info, "metadata": { "request_time_ms": round(total_elapsed_ms, 2), "server_url": state.server_url, "connection_statistics": state.statistics, }, } except ConnectionError as e: elapsed_ms = (time.perf_counter() - start_time) * 1000 # User-facing error update await ctx.error(f"Not connected when getting prompt '{name}': {str(e)}") # Detailed technical log logger.error( f"Not connected when getting prompt '{name}': {str(e)}", extra={"prompt_name": name, "duration_ms": elapsed_ms}, ) return { "success": False, "error": { "error_type": "not_connected", "message": str(e), "details": {"prompt_name": name}, "suggestion": "Use connect_to_server() to establish a connection first", }, "prompt": None, "metadata": { "request_time_ms": round(elapsed_ms, 2), }, } except Exception as e: elapsed_ms = (time.perf_counter() - start_time) * 1000 # Determine error type based on exception message error_type = "execution_error" suggestion = "Check the prompt name and arguments, then retry" error_msg = str(e).lower() if "not found" in error_msg or "unknown prompt" in error_msg or "no prompt" in error_msg: error_type = "prompt_not_found" suggestion = f"Prompt '{name}' does not exist on the server. Use list_prompts() to see available prompts" elif "argument" in error_msg or "parameter" in error_msg or "validation" in error_msg or "required" in error_msg: error_type = "invalid_arguments" suggestion = f"Arguments do not match the prompt schema. Use list_prompts() to see the correct schema for '{name}'" # User-facing error update await ctx.error(f"Failed to get prompt '{name}': {str(e)}") # Detailed technical log logger.error( f"Failed to get prompt '{name}': {str(e)}", extra={ "prompt_name": name, "arguments": arguments, "error_type": error_type, "duration_ms": elapsed_ms, }, ) # Increment error counter ConnectionManager.increment_stat("errors") return { "success": False, "error": { "error_type": error_type, "message": f"Failed to get prompt '{name}': {str(e)}", "details": { "prompt_name": name, "arguments": arguments, "exception_type": type(e).__name__, }, "suggestion": suggestion, }, "prompt": None, "metadata": { "request_time_ms": round(elapsed_ms, 2), }, }

Implementation Reference

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/rdwj/mcp-test-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server