Model Context Protocol moderate risk

Apache Airflow

This MCP server provides a standardized interface to Apache Airflow's REST API, enabling interaction with DAGs, tasks, variables, connections, and other Airflow components.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwriteexecadmin

Quickstart

Install

npx -y @smithery/cli install @yangkyeongmo/mcp-server-apache-airflow --client claude

Config

{
  "mcpServers": {
    "mcp-server-apache-airflow": {
      "command": "uvx",
      "args": ["mcp-server-apache-airflow"],
      "env": {
        "AIRFLOW_HOST": "https://your-airflow-host",
        "AIRFLOW_USERNAME": "your-username",
        "AIRFLOW_PASSWORD": "your-password"
      }
    }
  }
}

Exposed MCP Tools (5)

safe

List DAGs

Retrieves a list of available DAGs in Airflow.

Read-only operation, no side effects.

moderate

Create DAG Run

Triggers a new DAG run for a specified DAG.

Creates a new DAG run, potentially consuming resources.

high

Update Connection

Modifies an existing connection in Airflow.

Modifies connection details, potentially affecting access to external systems.

critical

Delete DAG

Removes a DAG from Airflow.

Destructive operation, permanently removes the DAG.

high

Set Task Instances State

Updates the state of specific task instances within a DAG run.

Can alter the execution flow and outcome of a DAG run.

Safety Assessment

This MCP server offers extensive control over Apache Airflow. It is safe to use for read-only operations and monitoring. However, caution is advised when performing write or delete operations, especially when dealing with sensitive data or critical workflows. Proper authentication and authorization are crucial to mitigate risks.

Wraps the official Apache Airflow REST API, ensuring compatibility.
Provides a standardized interface through the Model Context Protocol.
Supports a wide range of Airflow functionalities, allowing granular control.

Exposes Airflow's API, potentially allowing unauthorized access if not configured correctly.
Supports write and delete operations, which can be destructive.
Secrets management within Airflow connections and variables requires careful handling.
Lack of built-in sandboxing increases the risk of unintended consequences from actions.