1. Conduid
  2. Science
  3. Microscope MCP Client
MCP server · Science

Microscope MCP Client

MiCroscoPe is a lightweight command-line client that enables exhaustive testing of our MCP servers by AI agents.

37Low

Scored 6 months ago · breakdown

About Microscope MCP Client

Microscope MCP Client is an MCP server in the Science category: miCroscoPe is a lightweight command-line client that enables exhaustive testing of our MCP servers by AI agents.

Install

Claude Code
claude mcp add microscope-mcp-client -- npx -y microscope-mcp-client
npx
npx -y microscope-mcp-client

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Microscope MCP Client

Powered by Claude · Grounded in docs

I know everything about Microscope MCP Client. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • ✓README presentThe repository ships documentation.

5 checks have not been run yet.

README

MiCroscoPe is a lightweight command-line client that enables exhaustive testing of our MCP servers by AI agents.


Usage

As a CLI command (interactive mode)

# Connect to a remote server via npx
microscope --server "npx:@modelcontextprotocol/server-everything"

# Connect to a local server with custom logging level
microscope --server ./server.js --log-level debug
microscope --server ./server.py --log-level info

# Connect to a remote HTTP server
microscope --server "http://localhost:3000/mcp"
microscope --server "https://example.com/mcp"

Available commands

  • list - List all available tools
  • describe <toolName> - Show detailed information about a tool
  • call <toolName> '<jsonArgs>' - Execute a tool with JSON arguments
  • setLoggingLevel <level> - Configure the logging level
  • resources - List all available resources
  • resource <uri> - Show information about a specific resource
  • help - Show this help
  • exit or quit - Close the client

One-shot mode (single tool execution)

To execute a single tool and exit immediately:

# Single tool execution with custom logging level
microscope --server ./server.js --log-level debug --call-tool "<toolName> {\"toolParam1\":\"toolParamValue1\", \"toolParam2\":\"toolParamValue2\"}"

One-shot mode features:

  • Only shows the JSON response from the tool (all other logs are suppressed)
  • In case of parsing error or tool failure, writes a short error message to stderr and exits with non-zero code
  • The --call-tool argument expects a quoted string containing the tool name followed by a JSON object with parameters
  • If --call-tool is present, it runs non-interactively and exits immediately

Query available tools list:

# List all available tools with custom logging level
microscope --server ./server.js --log-level info --list-tools

As a library for test scripts

The client can also be used as an imported library within another project:

import { TestMcpClient } from 'microscope';

async function exampleUsage() {
    const client = new TestMcpClient();

    try {
        // Connect to an MCP server
        const serverTarget = {
            kind: 'npx',
            pkg: '@modelcontextprotocol/server-everything',
            args: ['stdio'],
            npxArgs: ['-y']
        };

        await client.connect(serverTarget, { quiet: true });

        // List available tools
        const tools = client.getTools();
        console.log(`Found ${tools.length} tools`);

        // Describe a specific tool
        const toolInfo = client.describeTool('echo');
        console.log('Echo tool:', toolInfo);

        // Call a tool
        const result = await client.callTool('echo', { message: 'Hello World!' });
        console.log('Result:', result);

        // List resources
        const resources = client.getResources();
        console.log(`Found ${resources.length} resources`);

    } finally {
        await client.disconnect();
    }
}

Library API:

  • new TestMcpClient() - Creates a new client instance
  • client.connect(target, options) - Connects to the MCP server
  • client.disconnect() - Disconnects from the server
  • client.getTools() - Returns list of available tools
  • client.describeTool(name) - Returns information about a specific tool
  • client.callTool(name, args) - Calls a tool with arguments
  • client.getResources() - Returns list of available resources
  • client.getResource(uri) - Returns information about a specific resource
  • client.setLoggingLevel(level) - Configures the logging level
  • client.getHandshakeInfo() - Returns handshake information
  • client.verifyHandshake() - Verifies that the handshake has completed

Installation

From npm

# Install from npm
npm install microscope-mcp-client

# Use the CLI
npx microscope-mcp-client --server "npx:@modelcontextprotocol/server-everything"

From GitHub Packages

# Configure npm to use GitHub Packages for @trevsmart scope
npm config set @trevsmart:registry https://npm.pkg.github.com

# Install the package
npm install @trevsmart/microscope-mcp-client

# Use the CLI
npx @trevsmart/microscope-mcp-client --server "npx:@modelcontextprotocol/server-everything"

License

ISC

README copy captured when this server was indexed; it may lag the repository. The original is authoritative.

Questions

About Microscope MCP Client

How do I install Microscope MCP Client?

Run claude mcp add microscope-mcp-client -- npx -y microscope-mcp-client, then add the server to your MCP client's configuration.

Is Microscope MCP Client safe to use with an AI agent?

Its trust score is 37 out of 100 (low). It passes 1 of the 1 static security checks run so far. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Microscope MCP Client still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.