MCP Workflow Orchestration Server

MCP Workflow Orchestration Server is an MCP server that provides storage, discovery, retrieval, and management of multi-step AI workflow playbooks defined in YAML files. Built on the @cyanheads/mcp-ts-core framework, it connects AI models and autonomous agents directly to a structured local workflow library without requiring external API keys. Developers and engineering teams use this server to standardize agent operations across complex multi-step processes, including CI/CD pipelines, repository maintenance, code reviews, and issue triage. The server indexes YAML workflow files organized by categories and maintains an in-memory semver-aware registry that automatically refreshes via filesystem watchers. Through standardized Model Context Protocol tools, agents can dynamically query available playbooks by tag or keyword, read full execution steps alongside global instruction files, create permanent or temporary workflows, and cleanly identify external tool dependencies. The architecture supports both standard input/output (stdio) and streamable HTTP transports, making it adaptable to local desktop assistants, containerized runners, and remote agent clusters.

Category: Developer Tools & Code Intelligence

Tags: automation, orchestration, task-runner, workflow, yaml

Visit MCP Workflow Orchestration Server

How to install and configure MCP Workflow Orchestration Server

To install and configure MCP Workflow Orchestration Server for your MCP client: 1. Ensure you have Bun v1.3.0 or higher (or Node.js v24+) installed on your machine. 2. Open your client configuration file, such as claude_desktop_config.json for Claude Desktop. 3. Add the server entry under mcpServers using bunx or npx: json { "mcpServers": { "workflows-mcp-server": { "type": "stdio", "command": "npx", "args": ["-y", "@cyanheads/workflows-mcp-server@latest"], "env": { "MCP_TRANSPORT_TYPE": "stdio", "WORKFLOWS_DIR": "/absolute/path/to/your/workflows-yaml" } } } } 4. Replace /absolute/path/to/your/workflows-yaml with the absolute path to your YAML playbooks directory. 5. Restart your MCP client to load the tools.

What you can do with MCP Workflow Orchestration Server

• Discovering existing team workflows by querying category names, text tags, or specific MCP server and tool dependencies. - Retrieving full YAML execution steps alongside global markdown instructions for prompt-driven execution in Claude Desktop or Cursor. - Generating and publishing reusable multi-step automation workflows directly into the permanent file library with automated semver tracking. - Creating throwaway session-specific workflows in a temporary directory for multi-stage exploratory coding or transient agent plans. - Deleting outdated or superseded workflow definitions from disk to maintain clean repository procedures across development teams.

Key facts

• https://github.com/cyanheads/workflows-mcp-server
• Developer Tools & Code Intelligence, Project & Task Management
• automation, orchestration, task-runner, workflow, yaml

Part of MCP Servers

What is MCP Workflow Orchestration Server?

MCP Workflow Orchestration Server is a TypeScript-based tool that lets AI agents discover, read, create, and manage YAML workflow files. It maintains an in-memory index of workflows and exposes five dedicated tools over STDIO or Streamable HTTP transports for autonomous agent orchestration.

How do I install MCP Workflow Orchestration Server?

You can configure it directly inside your MCP client settings using npx or bunx without a prior build step. Point the server to a local directory of YAML workflows using the WORKFLOWS_DIR environment variable, or run it via the official Docker container from GitHub Container Registry.

Which MCP clients work with MCP Workflow Orchestration Server?

It functions with any MCP-compliant client that supports standard input/output or HTTP transports. Tested environments include Claude Desktop, Cursor, and Visual Studio Code through their native extension and MCP configuration files.

What tools are included with MCP Workflow Orchestration Server?

The server provides five primary tools: workflow_list to browse and filter stored workflows, workflow_get to fetch step definitions and global instructions, workflow_create to save permanent versioned workflows, workflow_create_temp for throwaway playbooks, and workflow_delete to remove obsolete workflows.

Is MCP Workflow Orchestration Server open source?

Yes, MCP Workflow Orchestration Server is open source software released under the Apache 2.0 license. The source code and version history are available on GitHub at cyanheads/workflows-mcp-server.