Model Context Protocol moderate risk

Agentql

AgentQL MCP server enables structured data extraction from web pages using the AgentQL API, accessible via various MCP-compatible clients.

Connections & Capabilities

Connects To

GitHub

Capabilities

read

Quickstart

Install

npm install -g agentql-mcp

Config

{
  "mcp": {
    "inputs": [
      {
        "type": "promptString",
        "id": "apiKey",
        "description": "AgentQL API Key",
        "password": true
      }
    ],
    "servers": {
      "agentql": {
        "command": "npx",
        "args": ["-y", "agentql-mcp"],
        "env": {
          "AGENTQL_API_KEY": "${input:apiKey}"
        }
      }
    }
  }
}

Exposed MCP Tools (1)

moderate

extract-web-data

Extracts structured data from a specified URL based on a provided prompt describing the desired data fields.

While read-only, poorly crafted prompts could lead to unintended data extraction.

Safety Assessment

The AgentQL MCP server is relatively safe for data extraction tasks, assuming the AgentQL API is secure and the API key is properly managed. Risks are elevated if the API key is exposed or the prompt allows for unintended data retrieval.

Requires an API key for authentication.
Data extraction is limited to the specified URL and prompt.
No direct file system access or shell execution.
The server itself primarily performs data retrieval.

Relies on the security of the AgentQL API and its data extraction capabilities.
Potential for information leakage if the prompt is poorly designed or the API is compromised.
Exposure of the API key if not properly managed.
Over-reliance on external service availability.