diegosouzapw Praxis is a Command & Control (C2) framework for orchestrating AI coding agents. It provides a unified interface to manage, monitor, and interact with AI agents (like Claude Code, Cursor, Windsurf, etc.) running on remote machines.
First Step
Before using any commands, discover the full capabilities by running:
praxis_cli --fullhelp
This outputs comprehensive documentation for all commands and subcommands.
Key Concepts
- Node: A machine running the Praxis node agent
- Agent: An AI coding agent (e.g., Claude Code) discovered on a node
- Session: An active connection to an agent for sending prompts
- Operation: A pre-configured prompt/workflow for common tasks
- Chain: A sequence of operations executed as a workflow
Requirements
The Praxis service must be running and accessible via RabbitMQ. The default connection is amqp://praxis:praxis@localhost:5672.
To specify a different RabbitMQ URL:
praxis_cli --rabbitmq amqp://user:pass@host:5672 node list
Or set the environment variable:
export PRAXIS_RABBITMQ_URL=amqp://user:pass@host:5672
Output Formats
Use --output json for machine-readable output suitable for scripting and parsing.