Iron Manus MCP

A modular MCP server for task orchestration, API integration, and knowledge synthesis using a finite state machine.

Model Context Protocol server for AI workflow orchestrationinspired by Manus AI's orchestration patterns and Andrej Karpathy's "Iron Man" analogy.

Extended context management through agent delegation and Todos as a custom agent dispatch tool

Historical Notes: Architectural Patterns

Note: This project is now archived. The following notes document architectural decisions that later appeared in mainstream AI tooling.

Several patterns implemented in Iron Manus MCP (June 2024) were later adopted by Claude Code and similar tools. These emerged from independent experimentation rather than foresight, solving problems that turned out to be common across the ecosystem.

interface TodoItem { type?: 'TaskAgent' | 'SubAgent' | 'DirectExecution'; meta_prompt?: MetaPrompt; // Declarative agent configuration }

This MCP was the first to use todos as a subagent dispatch queue. Claude Code had TodoWrite/TodoRead at that point, but nobody (including Anthropic) was thinking of todos as an agent coordination primitive. The Task tool existed but was just non-specialized agents. Our implementation connected the two: todos become task dispatches. Boris Cherny announced "We're turning Todos into Tasks" on January 22, 2026, seven months after we shipped the same idea.

const PHASE_ALLOWED_TOOLS = { PLAN: ['TodoWrite'], EXECUTE: ['TodoRead', 'TodoWrite', 'Task', 'Bash', 'Read', 'Write', 'Edit'], VERIFY: ['TodoRead', 'Read'], // Read-only during verification };

Restricting tool availability by phase prevented misuse (e.g., writing files during verification). This pattern appears in Claude Code Agent Teams (February 2026) via phase-based permissions.

The explicitINIT → QUERY → ENHANCE → KNOWLEDGE → PLAN → EXECUTE → VERIFY → DONEworkflow enforced planning before execution. Claude Code's Plan Mode (August 2025) provides similar structure.

4. Context Isolation via File-Based Communication

./iron-manus-sessions/{session_id}/ ├── synthesized_knowledge.md ├── primary_research.md └── agent_output.md

Task()agents have isolated contexts and cannot share state directly. This project used session workspaces for inter-agent coordination. Claude Code Agent Teams implements similar patterns via~/.claude/teams/and~/.claude/tasks/.

Nine specialized roles (planner,coder,critic,researcher,analyzer,synthesizer,ui_architect,ui_implementer,ui_refiner) with distinct thinking methodologies. Claude Code custom subagents (July 2025) provide similar specialization.

(ROLE: coder) (CONTEXT: auth_system) (PROMPT: Implement JWT auth) (OUTPUT: auth_module.ts)

Declarative syntax for agent configuration embedded in todo content. Similar patterns appear in Claude Code's subagent configuration.

Iron Manus MCP (June 2024) introduced todos as subagent dispatch, phase-gated tools, structured planning phases, context isolation, and role-based agents. Claude Code adopted these patterns between July 2025 and February 2026. These patterns emerged from practical necessity. Multi-agent orchestration requires task decomposition, context isolation, and workflow structure, all of which eventually appeared in production tooling.

8-phase workflow orchestration:INIT → QUERY → ENHANCE → KNOWLEDGE → PLAN → EXECUTE → VERIFY → DONE

- JARVIS- 8-phase workflow controller
- APITaskAgent- API discovery and fetching with SSRF protection
- PythonComputationalTool- Python execution for data analysis
- IronManusStateGraph- Session state management
- SlideGenerator- HTML slide generation
- HealthCheck- Runtime diagnostics

git clone https://github.com/dnnyngyen/iron-manus-mcp cd iron-manus-mcp npm install npm run build npm start

docker build -t iron-manus-mcp . docker run -d --name iron-manus-mcp iron-manus-mcp

claude mcp add iron-manus-mcp node dist/index.js

{ "mcpServers": { "iron-manus-mcp": { "command": "node", "args": ["path/to/iron-manus-mcp/dist/index.js"] } } }

ALLOWED_HOSTS=api.github.com,httpbin.org # SSRF whitelist ENABLE_SSRF_PROTECTION=true # Enable security KNOWLEDGE_MAX_CONCURRENCY=2 # API concurrency limit KNOWLEDGE_TIMEOUT_MS=4000 # Request timeout (ms)

npm run build # Compile TypeScript npm run lint # Check code style npm run format # Format code npm start # Run server npm run dev # Build + watch mode

- SSRF protection blocks private IPs (192.168.x.x, 127.x.x.x, etc.)
- URL validation (HTTP/HTTPS only)
- Host allowlist enforcement
- Request timeout and size limits

Magica is your all-in-one AI platform, offering 2500+ cutting-edge tools under a single subscription.

Connect to the Taskade platform via MCP. Access tasks, projects, workflows, and AI agents in real-time through a unified workspace and API.

Dynamically search and call tools using UnifAI Network

A multi-client AI agent monitoring and control system with automatic task completion detection.

An MCP server for AI-to-AI collaboration, enabling autonomous workflows and role-based task management between different AI models.

Connect your GTD system directly to any LLM, so you can capture, organize, and review your life and work using natural language.

Unified MCP gateway that loads 150+ AI tools into Claude Desktop via a single Docker command on Windows. Routes tool calls across specialized servers (Business Intelligence, API Hub, Content Automation, Intelligence) with centralized logging and governance. Built for teams without dedicated platform engineers.

About AI-powered Jira CLI and MCP server for humans and agents manage issues, sprints, boards with interactive wizards, multi-provider AI

A Python monorepo for AI-powered project management and productivity servers, utilizing the Claude API.

Manages AI agent handoffs with structured documentation and seamless task transitions.