iron-manus
Iron Manus MCP
claude mcp add --transport stdio dnnyngyen-iron-manus-mcp node path/to/iron-manus-mcp/dist/index.js \ --env ALLOWED_HOSTS="api.github.com,httpbin.org # SSRF whitelist" \ --env KNOWLEDGE_TIMEOUT_MS="4000 # Request timeout (ms)" \ --env ENABLE_SSRF_PROTECTION="true # Enable security" \ --env KNOWLEDGE_MAX_CONCURRENCY="2 # API concurrency limit"
How to use
Iron Manus MCP is an 8-phase workflow orchestration server designed to manage AI-driven tasks across multiple subagents and tools. It exposes a Node.js-based MCP endpoint that coordinates planning, execution, and verification of tasks through a structured Todo/agent system. You can integrate it into your Claude Code or generic MCP workflow by pointing your MCP config to run the server’s entry point, and you can optionally tune safety settings like SSRF protection and request timeouts to suit your deployment needs. The server exposes a set of tools and patterns (phase-gated tool access, session-based context isolation, and role-based subagents) that help you orchestrate complex AI workflows with clear separation of planning, execution, and validation phases. Tools mentioned in the project include an 8-phase workflow controller, API discovery, Python-based computation, session state management, slide generation, and runtime health checks.
To use it, install dependencies, build, and start the server locally. You can then integrate it into your MCP setup by referencing the server in your mcp_config, for example using a node command that runs dist/index.js. Ensure you configure environment variables for SSRF protection, allowed hosts, and knowledge concurrency/timeouts to match your security and performance needs.
How to install
Prerequisites:
- Node.js 20+ and npm (as recommended by the project)
- Git (optional for cloning the repository)
From source (recommended):
-
Clone the repository git clone https://github.com/dnnynguyen/iron-manus-mcp.git cd iron-manus-mcp
-
Install dependencies npm install
-
Build the project npm run build
-
Run the MCP server npm start
Docker (alternative):
-
Build the image docker build -t iron-manus-mcp .
-
Run the container docker run -d --name iron-manus-mcp iron-manus-mcp
Or with docker-compose (if provided in repo): docker-compose up -d
MCP configuration integration (example):
- Add to Claude Code or your MCP config: { "mcpServers": { "iron-manus-mcp": { "command": "node", "args": ["path/to/iron-manus-mcp/dist/index.js"] } } }
Environment variables you may configure:
- ALLOWED_HOSTS: Comma-separated list of allowed hosts for SSRF protection
- ENABLE_SSRF_PROTECTION: true/false to enable SSRF protections
- KNOWLEDGE_MAX_CONCURRENCY: Maximum concurrent knowledge/API calls
- KNOWLEDGE_TIMEOUT_MS: Timeout for knowledge/API requests in milliseconds
Additional notes
Tips and considerations:
- This MCP server emphasizes safety with SSRF protection and host allowlisting. Set ALLOWED_HOSTS and ENABLE_SSRF_PROTECTION according to your network.
- The project uses a session-based workspace pattern to isolate agent contexts; avoid sharing session directories across concurrent tasks.
- When deploying in production, consider enabling strict timeouts and concurrency limits to prevent runaway tasks.
- The MCP integration example shows how to reference the server in your MCP configuration. If you modify path names or build outputs, ensure the node command points to the correct dist/index.js file.
- If you encounter build or runtime issues, consult the repository’s security notes and ensure you’re running compatible Node.js versions (as indicated by the Node.js badge in the README).
Related MCP Servers
context7
Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors
cclsp
Claude Code LSP: enhance your Claude Code experience with non-IDE dependent LSP integration.
systemprompt-code-orchestrator
MCP server for orchestrating AI coding agents (Claude Code CLI & Gemini CLI). Features task management, process execution, Git integration, and dynamic resource discovery. Full TypeScript implementation with Docker support and Cloudflare Tunnel integration.
airtable
Airtable integration for AI-powered applications via Anthropic's Model Context Protocol (MCP)
modular
A Model Context Protocol (MCP) proxy server that enables efficient management of large tool collections across multiple MCP servers by grouping them and loading tool schemas on-demand.
ZMCPTools
A custom TypeScript MCP Server intended to be used with Claude Code