Omnispindle - Advanced MCP Todo Management System

Name: Omnispindle - Advanced MCP Todo Management System
Author: MadnessEngineering

@MadnessEngineering

About Omnispindle - Advanced MCP Todo Management System

A comprehensive MCP-based todo management system, that serves as a central nervous system for Madness Interactive, a multi-project task coordination workshop.

#mcp-server

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "omnispindle-madnessengineering": {
      "command": "python",
      "args": [
        "-m",
        "src.Omnispindle.stdio_server"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Omnispindle - Advanced MCP Todo Management System?

Omnispindle is a Python FastMCP server with 38 tools that provides AI agents with a unified interface for task management, knowledge capture, session tracking, epic goal planning, and context retrieval. It is part of the Madness Interactive ecosystem, packaged on PyPI, integrated with Auth0, and works wherever an MCP configuration exists.

How to use Omnispindle - Advanced MCP Todo Management System?

Install via pip install omnispindle, then run the CLI command omnispindle-stdio for Claude Desktop or omnispindle for an HTTP web server. Add the server configuration to claude_desktop_config.json with environment variables for mode and tool loadout. The first tool call opens a browser for Auth0 login; tokens are saved locally. Development setup involves cloning the repository, installing dependencies, and running python -m src.Omnispindle.stdio_server.

Key features of Omnispindle - Advanced MCP Todo Management System

38 MCP tools across seven categories for comprehensive task and knowledge management.
Tool loadouts control which tools are registered, optimizing token usage.
Zero-config Auth0 device flow for automatic authentication on first use.
Semantic search via vector embeddings for relevant todos and lessons.
Session genealogy allows tracing work through parent-child relationships.
Quest system for managing multi-step epic goals with progress tracking.

Use cases of Omnispindle - Advanced MCP Todo Management System

AI agents managing tasks with full metadata, priorities, and dependency tracking.
Persistent knowledge capture across projects, searchable by text or vector embedding.
Tracking AI work sessions with fork, spawn, and lineage capabilities.
Coordinating cross-agent efforts via shared journals and context bundles.
Running a lightweight todo system with minimal loadout for token‑constrained environments.

FAQ from Omnispindle - Advanced MCP Todo Management System

What are tool loadouts and how do I choose one?

Tool loadouts (OMNISPINDLE_TOOL_LOADOUT) control the number of registered tools—from full (38) to minimal (4). Choose based on your agent's token budget and required functionality; e.g., basic provides 15 tools for core todo CRUD, context, and quests.

How does authentication work?

Omnispindle uses zero‑config Auth0 device flow. On the first tool call, a browser opens for login. The token saves locally, and all subsequent calls are authenticated automatically. For manual setup, you can run python -m src.Omnispindle.token_exchange or set environment variables.

What operation modes are available and when should I use each?

Set OMNISPINDLE_MODE to api (HTTP calls to the ecosystem API), hybrid (default; API with MongoDB fallback), local (direct MongoDB), or auto (benchmarks and picks the fastest). Use api for cloud‑native setups, local for offline development.

How is data isolated and where does it live?

All data is scoped per user via the MCP_USER_EMAIL environment variable. The database is MongoDB (database swarmonomicon). In API mode, data goes through madnessinteractive.cc/api; in local/hybrid mode, you connect your own MongoDB instance.

What transport and auth does the server support?

The server supports stdio transport (for Claude Desktop) via omnispindle-stdio and an HTTP web server via omnispindle. Authentication uses Auth0 device flow; tokens can also be provided through MADNESS_AUTH_TOKEN or AUTH0_TOKEN environment variables.