@alpic-ai/grizzly v0.3.0

Inspector Grizzly

The Inspector Grizzly is based on the MCP Inspector by anthropic!

What's new ?

Grizzly packages a bunch of additional features :

• A server analyzer to scan unknown MCP server and check for security issue (like Prompt injection and Poisoned-tools)
• A playground to use MCP server tools within the context of a chat conversation and avoid going back and forth between the inspector and an LLM application
• A suite of good practice tests on MCP server tools to help you improve your MCP server discoverability
• A test case generator to assert potential tool confusion among your MCP server capabilities

Running the Inspector

Requirements

• Node.js: ^22.7.5

From an MCP server repository

To inspect an MCP server implementation, there's no need to clone this repo. Instead, use npx. For example, if your server is built at build/index.js:

npx @alpic-ai/grizzly node build/index.js

You can pass both arguments and environment variables to your MCP server. Arguments are passed directly to your server, while environment variables can be set using the -e flag:

# Pass arguments only
npx @alpic-ai/grizzly node build/index.js arg1 arg2

# Pass environment variables only
npx @alpic-ai/grizzly -e key=value -e key2=$VALUE2 node build/index.js

# Pass both environment variables and arguments
npx @alpic-ai/grizzly -e key=value -e key2=$VALUE2 node build/index.js arg1 arg2

# Use -- to separate inspector flags from server arguments
npx @alpic-ai/grizzly -e key=$VALUE -- node build/index.js -e server-flag

The inspector runs both an MCP Inspector (MCPI) client UI (default port 6274) and an MCP Proxy (MCPP) server (default port 6277). Open the MCPI client UI in your browser to use the inspector. (These ports are derived from the T9 dialpad mapping of MCPI and MCPP respectively, as a mnemonic). You can customize the ports if needed:

CLIENT_PORT=8080 SERVER_PORT=9000 npx @alpic-ai/grizzly node build/index.js

For more details on ways to use the inspector, see the Inspector section of the MCP docs site. For help with debugging, see the Debugging guide.

Servers File Export

The MCP Inspector provides convenient buttons to export server launch configurations for use in clients such as Cursor, Claude Code, or the Inspector's CLI. The file is usually called mcp.json.

• Server Entry - Copies a single server configuration entry to your clipboard. This can be added to your mcp.json file inside the mcpServers object with your preferred server name.

  STDIO transport example:

  {
    "command": "node",
    "args": ["build/index.js", "--debug"],
    "env": {
      "API_KEY": "your-api-key",
      "DEBUG": "true"
    }
  }

  SSE transport example:

  {
    "type": "sse",
    "url": "http://localhost:3000/events",
    "note": "For SSE connections, add this URL directly in Client"
  }

• Servers File - Copies a complete MCP configuration file structure to your clipboard, with your current server configuration added as default-server. This can be saved directly as mcp.json.

  STDIO transport example:

  {
    "mcpServers": {
      "default-server": {
        "command": "node",
        "args": ["build/index.js", "--debug"],
        "env": {
          "API_KEY": "your-api-key",
          "DEBUG": "true"
        }
      }
    }
  }

  SSE transport example:

  {
    "mcpServers": {
      "default-server": {
        "type": "sse",
        "url": "http://localhost:3000/events",
        "note": "For SSE connections, add this URL directly in Client"
      }
    }
  }

These buttons appear in the Inspector UI after you've configured your server settings, making it easy to save and reuse your configurations.

For SSE transport connections, the Inspector provides similar functionality for both buttons. The "Server Entry" button copies the SSE URL configuration that can be added to your existing configuration file, while the "Servers File" button creates a complete configuration file containing the SSE URL for direct use in clients.

You can paste the Server Entry into your existing mcp.json file under your chosen server name, or use the complete Servers File payload to create a new configuration file.

Authentication

The inspector supports bearer token authentication for SSE connections. Enter your token in the UI when connecting to an MCP server, and it will be sent in the Authorization header. You can override the header name using the input field in the sidebar.

Security Considerations

The MCP Inspector includes a proxy server that can run and communicate with local MCP processes. The proxy server should not be exposed to untrusted networks as it has permissions to spawn local processes and can connect to any specified MCP server.

Configuration

The MCP Inspector supports the following configuration settings. To change them, click on the Configuration button in the MCP Inspector UI:

These settings can be adjusted in real-time through the UI and will persist across sessions.

The inspector also supports configuration files to store settings for different MCP servers. This is useful when working with multiple servers or complex configurations:

npx @alpic-ai/grizzly --config path/to/config.json --server everything

Example server configuration file:

{
  "mcpServers": {
    "everything": {
      "command": "npx",
      "args": ["@modelcontextprotocol/server-everything"],
      "env": {
        "hello": "Hello MCP!"
      }
    },
    "my-server": {
      "command": "node",
      "args": ["build/index.js", "arg1", "arg2"],
      "env": {
        "key": "value",
        "key2": "value2"
      }
    }
  }
}

Tip: You can easily generate this configuration format using the Server Entry and Servers File buttons in the Inspector UI, as described in the Servers File Export section above.

You can also set the initial transport type, serverUrl, serverCommand, and serverArgs via query params, for example:

http://localhost:6274/?transport=sse&serverUrl=http://localhost:8787/sse
http://localhost:6274/?transport=streamable-http&serverUrl=http://localhost:8787/mcp
http://localhost:6274/?transport=stdio&serverCommand=npx&serverArgs=arg1%20arg2

You can also set initial config settings via query params, for example:

http://localhost:6274/?MCP_SERVER_REQUEST_TIMEOUT=10000&MCP_REQUEST_TIMEOUT_RESET_ON_PROGRESS=false&MCP_PROXY_FULL_ADDRESS=http://10.1.1.22:5577

Note that if both the query param and the corresponding localStorage item are set, the query param will take precedence.

From this repository

If you're working on the inspector itself:

Development mode:

npm run dev

Note for Windows users: On Windows, use the following command instead:

npm run dev:windows

Production mode:

npm run build
npm start

CLI Mode

CLI mode enables programmatic interaction with MCP servers from the command line, ideal for scripting, automation, and integration with coding assistants. This creates an efficient feedback loop for MCP server development.

npx @alpic-ai/grizzly --cli node build/index.js

The CLI mode supports most operations across tools, resources, and prompts. A few examples:

# Basic usage
npx @alpic-ai/grizzly --cli node build/index.js

# With config file
npx @alpic-ai/grizzly --cli --config path/to/config.json --server myserver

# List available tools
npx @alpic-ai/grizzly --cli node build/index.js --method tools/list

# Call a specific tool
npx @alpic-ai/grizzly --cli node build/index.js --method tools/call --tool-name mytool --tool-arg key=value --tool-arg another=value2

# List available resources
npx @alpic-ai/grizzly --cli node build/index.js --method resources/list

# List available prompts
npx @alpic-ai/grizzly --cli node build/index.js --method prompts/list

# Connect to a remote MCP server
npx @alpic-ai/grizzly --cli https://my-mcp-server.example.com

# Call a tool on a remote server
npx @alpic-ai/grizzly --cli https://my-mcp-server.example.com --method tools/call --tool-name remotetool --tool-arg param=value

# List resources from a remote server
npx @alpic-ai/grizzly --cli https://my-mcp-server.example.com --method resources/list

UI Mode vs CLI Mode: When to Use Each

License

This project is licensed under the MIT License—see the LICENSE file for details.