Omnispindle
by MadnessEngineering
Overview
Omnispindle acts as a coordination layer providing standardized Model Context Protocol (MCP) tools for AI agents to manage todos, capture lessons, and facilitate cross-project coordination within the Madness Interactive ecosystem.
Installation
docker run -e OMNISPINDLE_MODE=api omnispindle:v1.0.0Environment Variables
- OMNISPINDLE_MODE
- OMNISPINDLE_TOOL_LOADOUT
- OMNISPINDLE_FALLBACK_ENABLED
- MADNESS_API_URL
- AUTH0_TOKEN
- MADNESS_AUTH_TOKEN
- MADNESS_API_KEY
- MCP_USER_EMAIL
- MCP_USER_ID
- MONGODB_URI
- MONGODB_DB
- MQTT_HOST
- MQTT_PORT
- HOST
- PORT
- AUTH0_DOMAIN
- AUTH0_CLIENT_ID
- AUTH0_AUDIENCE
Security Notes
CRITICAL: The `bring_your_own` tool (in `src/Omnispindle/tools.py`) allows arbitrary code execution (Python, JavaScript, Bash) via `subprocess.run` and `importlib.util.module_from_spec` (equivalent to `eval`). While its docstring claims it is 'DISABLED for security reasons,' the code for execution is present and it is enabled in the default 'full' tool loadout when running in local mode. This is a severe security risk. Additionally, Auth0 configuration defaults (domain, client ID, audience) are hardcoded in `src/Omnispindle/auth_utils.py` and `src/Omnispindle/token_exchange.py`, which the README explicitly warns against deploying as-is. The `mqtt_publish` function uses `subprocess.run` directly with input strings, posing a potential command injection risk if not properly sanitized.
Similar Servers
trello-mcp-server
Provides an AI-agent-friendly interface (MCP server) to interact with the Trello API for managing boards, lists, cards, and other Trello resources.
Teamwork-MCP
The MCP server simplifies interaction with the Teamwork.com API, enabling AI agents to manage projects, tasks, people, and companies within Teamwork.
AI-Prompt-Guide-MCP
Orchestrates AI agents for project management and development workflows by linking structured markdown specifications and tasks.
todo-mcp-server
Provides autonomous task management and random string generation for AI agents via Model Context Protocol (MCP) using stdio or HTTP transports.