centaur
NewMultiplayer, self-hosted, secure agents.
Summary
Centaur enables multiplayer, self-hosted, secure agent workflows within Claude Code, allowing multiple users to collaborate on complex tasks with isolated, safe execution environments.
- It is useful for teams needing to run automated agents in a controlled, private setup without relying on external services.
Install & Usage
mkdir -p .claude/agentsAdd the configuration to .claude/agents/centaur.md
@centaurUse Cases
Usage Examples
/centaur spawn agent1 --task 'Research the latest Python 3.13 features and summarize them in bullet points'
/centaur run --multi --agents agent1,agent2 --workflow 'agent1: fetch data from API; agent2: parse and save to CSV'
/centaur secure --agent myagent --command 'Run all unit tests and report failures'
Security Audits
Frequently Asked Questions
What is centaur?
Centaur enables multiplayer, self-hosted, secure agent workflows within Claude Code, allowing multiple users to collaborate on complex tasks with isolated, safe execution environments. It is useful for teams needing to run automated agents in a controlled, private setup without relying on external services.
How to install centaur?
To install centaur: create the agents directory (mkdir -p .claude/agents), then add the config to .claude/agents/centaur.md. Finally, @centaur in Claude Code.
What is centaur best for?
centaur is a agent categorized under General. It is designed for: agent. Created by paradigmxyz.
What can I use centaur for?
centaur is useful for: Run a multi-step research task where one agent gathers data and another analyzes it, all within a secure self-hosted environment.; Collaborate with team members on a shared codebase by spawning isolated agents that can safely test and modify code without conflicts.; Automate CI/CD pipeline steps using self-hosted agents that execute build and deployment scripts securely.; Deploy a fleet of agents to monitor system logs and trigger alerts based on predefined patterns, all managed locally.; Simulate multi-agent conversations for testing chatbot interactions in a private, reproducible setting.; Offload long-running computations to a background agent while continuing to work on other tasks in Claude Code..