mwaa-mcp-server

mwaa-mcp-server

MCP server for Amazon MWAA that enables managing Airflow environments, DAGs, runs, task instances, and configurations through natural language, using secure AWS-native API access and read-only by default.

Category
访问服务器

README

mwaa-mcp-server

A Model Context Protocol (MCP) server for Amazon Managed Workflows for Apache Airflow (MWAA).

Features

  • MWAA Environment Management: List, describe, create, update, and delete MWAA environments
  • Airflow DAG Operations: List DAGs, view details, get source code, trigger runs, pause/unpause
  • DAG Run Monitoring: List and inspect DAG runs, task instances, and task logs
  • Configuration Access: View Airflow connections (with password redaction) and variables
  • Error Diagnostics: Retrieve DAG import/parsing errors
  • Secure by Design: All Airflow API operations go through AWS invoke_rest_api — no CLI or web tokens are exposed
  • Read-only by Default: Write operations require the --allow-write flag
  • Airflow 2.x and 3.x: The server detects each environment's Airflow major version and translates API differences (e.g. execution_date vs logical_date) transparently — tools keep one stable interface

Prerequisites

  • Python 3.10+
  • AWS credentials configured (via environment variables, AWS CLI profile, or IAM role)
  • An Amazon MWAA environment (Airflow 2.x or 3.x)

Installation

Using uvx (recommended)

uvx --from git+https://github.com/biswasbiplob/mwaa-mcp-server mwaa-mcp-server

Using pip

pip install git+https://github.com/biswasbiplob/mwaa-mcp-server

Using Docker

docker build -t mwaa-mcp-server .
docker run -it mwaa-mcp-server

Configuration

Environment Variables

Variable Description Default
AWS_REGION AWS region for API calls us-east-1
AWS_PROFILE AWS CLI profile name Default credential chain
MWAA_ENVIRONMENT Default MWAA environment name (avoids per-call specification) Auto-detect
FASTMCP_LOG_LEVEL Log level (ERROR, WARNING, INFO, DEBUG) WARNING

CLI Arguments

Argument Description
--allow-write Enable write operations (trigger DAG runs, pause/unpause DAGs, create/update/delete environments)

Usage with Claude Desktop

Add to your Claude Desktop MCP configuration:

{
  "mcpServers": {
    "mwaa-mcp-server": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/biswasbiplob/mwaa-mcp-server", "mwaa-mcp-server"],
      "env": {
        "AWS_REGION": "us-east-1",
        "AWS_PROFILE": "my-profile",
        "MWAA_ENVIRONMENT": "my-environment",
        "FASTMCP_LOG_LEVEL": "INFO"
      }
    }
  }
}

For write access:

{
  "mcpServers": {
    "mwaa-mcp-server": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/biswasbiplob/mwaa-mcp-server", "mwaa-mcp-server", "--allow-write"],
      "env": {
        "AWS_REGION": "us-east-1",
        "AWS_PROFILE": "my-profile",
        "MWAA_ENVIRONMENT": "my-environment",
        "FASTMCP_LOG_LEVEL": "INFO"
      }
    }
  }
}

Tool Reference

Environment Management Tools

Tool Description Access
list-environments List all MWAA environments in the region Read
get-environment Get detailed information about an environment Read
create-environment Create a new MWAA environment Write
update-environment Update an existing MWAA environment Write
delete-environment Delete an MWAA environment Write

Airflow Tools

Tool Description Access
list-dags List all DAGs in an environment Read
get-dag Get details of a specific DAG Read
get-dag-source Get the source code of a DAG by DAG ID Read
list-dag-runs List DAG runs for a specific DAG Read
get-dag-run Get details of a specific DAG run Read
list-task-instances List task instances for a DAG run Read
get-task-instance Get details of a specific task instance (supports map_index for mapped tasks) Read
list-mapped-task-instances List all mapped instances of a dynamically mapped task Read
get-task-logs Get logs for a task instance try (supports pagination for large logs) Read
list-connections List Airflow connections (passwords redacted) Read
list-variables List Airflow variables (sensitive values redacted) Read
get-import-errors Get DAG import/parsing errors Read
trigger-dag-run Trigger a new DAG run Write
pause-dag Pause a DAG Write
unpause-dag Unpause a DAG Write
clear-task-instances Clear task instances for a DAG, allowing re-runs Write

IAM Permissions

Minimum Read-only Policy

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "airflow:GetEnvironment",
        "airflow:ListEnvironments",
        "airflow:InvokeRestApi"
      ],
      "Resource": "*"
    }
  ]
}

Full Access Policy (Read + Write)

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "airflow:GetEnvironment",
        "airflow:ListEnvironments",
        "airflow:CreateEnvironment",
        "airflow:UpdateEnvironment",
        "airflow:DeleteEnvironment",
        "airflow:InvokeRestApi"
      ],
      "Resource": "*"
    }
  ]
}

Security

  • No token exposure: Unlike other MWAA tools, this server does not expose create_cli_token or create_web_login_token as tools. All Airflow API operations go through the AWS-native invoke_rest_api.
  • Password redaction: Connection passwords and extra fields are automatically redacted in responses.
  • Variable redaction: Variable values with sensitive keys (matching patterns like secret, password, token, api_key) are automatically redacted.
  • Input validation: Environment names and path parameters are validated to prevent injection.
  • Read-only by default: Write operations are disabled unless --allow-write is explicitly passed.

License

Apache-2.0

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选