Which integrations are available for this server?

Enables AI agents to programmatically create, manage, and execute independent [Python](/mcp/servers/integrations/python) workflow scripts with full access to the Python ecosystem. Allows posting messages to [Slack](/mcp/servers/integrations/slack) channels via webhook URLs through executable workflow scripts.

Workflows MCP Server

License: MIT

A Model Context Protocol (MCP) server that enables AI agents to programmatically create, manage, and execute independent Python workflow scripts. This provides a powerful, flexible, and vendor-agnostic way to automate tasks using AI.

Unlike platforms like Make or Dify, which only allow AI to execute pre-built workflows, this server empowers AI to create and modify the workflows themselves as simple Python scripts.

Features

Programmatic Workflow Creation: AI can generate new Python workflow scripts from scratch.
Independent Scripts: Each workflow is a self-contained .py file, making them portable, versionable, and easy to debug.
Full Flexibility: Leverage the entire Python ecosystem within your workflows.
Simple Execution: Run any workflow by name with optional parameters.
Full CRUD Operations: Create, Read, Update, and Delete workflows via MCP tools.
Vendor-Agnostic: Built on the open Model Context Protocol standard.

Getting Started

Prerequisites

Python 3.10+
An MCP-compatible client (e.g., Manus, Cursor, LobeChat)

Installation

Clone the repository:
git clone https://github.com/YoruLabs/Workflows-MCP.git cd Workflows-MCP
Install dependencies:
pip install -e .
Run the server:
workflows-mcp
The server will start and be available for your MCP client to connect to.

Configuration

Workflows Directory: By default, workflows are stored in the workflows/ directory. You can change this by setting the WORKFLOWS_DIR environment variable.
Slack Webhook: For the example_slack_message workflow, set the SLACK_WEBHOOK_URL environment variable.

MCP Tools

The server exposes the following tools for AI agents:

Tool Name	Description
`create_workflow`	Creates a new Python workflow script from a name, description, and code.
`execute_workflow`	Executes a workflow by name, passing optional JSON parameters.
`list_workflows`	Lists all available workflow scripts with their metadata.
`read_workflow`	Reads the full source code of a specified workflow script.
`update_workflow`	Updates the description or code of an existing workflow.
`delete_workflow`	Deletes a workflow script from the filesystem.

Creating a Workflow

To create a workflow, the AI agent calls the create_workflow tool. The code parameter must contain a Python script with a run function.

Example

def run(params: dict = None) -> dict: """ This is the entry point for the workflow. Args: params: An optional dictionary of parameters passed during execution. Returns: A dictionary with the results of the workflow. """ params = params or {} name = params.get("name", "World") # ... your workflow logic ... return {"status": "success", "greeting": f"Hello, {name}!"}

Example Workflows

This repository includes two example workflows in the workflows/ directory:

example_hello_world.py: A simple workflow that takes a name and uppercase parameter and returns a greeting.
example_slack_message.py: A workflow that posts a message to a Slack channel using a webhook URL.