This project provides a framework for running autonomous AI agents, specifically Claude, for software development tasks. It targets developers and teams looking to automate complex, multi-stage projects by enabling agents to self-schedule, coordinate, and persist work across multiple codebases without constant human oversight.

How It Works

The orchestrator leverages tmux to manage multiple persistent terminal sessions, each housing an AI agent. It employs a three-tier hierarchy: an Orchestrator at the top, Project Managers in the middle, and individual Engineers at the bottom. This separation addresses context window limitations by assigning specialized roles and allows for parallel workstreams, with agents communicating via custom shell scripts (send-claude-message.sh, schedule_with_note.sh) that interact with tmux windows.

Quick Start & Requirements

• Install: No explicit installation instructions are provided beyond needing tmux and a way to run Claude models (e.g., via a CLI or API).
• Prerequisites: tmux, a Claude model accessible via terminal.
• Setup: Basic setup involves creating a project_spec.md and starting tmux sessions. Full setup involves creating multiple tmux sessions for orchestrator and project managers.

Highlighted Details

• Enables agents to self-schedule check-ins and coordinate tasks.
• Utilizes a hierarchical agent structure (Orchestrator, Project Manager, Engineer) to manage complexity.
• Automates Git commits every 30 minutes and tags stable versions.
• Provides scripts for simplified agent communication and scheduling.

Maintenance & Community

The project is community-driven, with contributions encouraged for new patterns, workflows, and optimizations. No specific community links (Discord, Slack) or contributor details are provided in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive for commercial use and integration with closed-source projects.

Limitations & Caveats

The effectiveness relies heavily on the quality of agent prompts and specifications. The setup requires familiarity with tmux and managing AI model interactions via the command line. The project is presented as a framework, implying that significant prompt engineering and script customization may be necessary for robust operation.