Overview

SwarmForge is a sophisticated, yet lightweight, tmux-based platform designed to orchestrate multiple AI agents, transforming them into a cohesive and professional software engineering team. It provides a disciplined framework for agent communication and collaboration across different Git worktrees, ensuring seamless project execution without conflicts.

Key Features

• Config-Driven Swarm Topology: Define your AI agent team structure and interactions through a flexible swarmforge.conf file, allowing for custom swarms tailored to specific project needs.
• Project-Local Roles & Layered Constitution: Each agent's behavior is dictated by project-local prompt files, layered with a flexible constitution for adaptable directives.
• Multiple Workflow Packs: Choose from pre-defined agent workflows (two-pack, four-pack, six-pack) or create your own, each optimized for different project complexities and quality gate requirements.
• Integrated Pack Cockpit Dashboard: Monitor and manage your agent swarm through an intuitive web dashboard, featuring task management, agent status, and direct communication channels.
• Seamless Handoff Protocol: Agents communicate via a robust, file-based handoff system that ensures data integrity and reliable task progression.
• Self-Hosted & Lightweight: Runs entirely locally, leveraging tmux and a web browser for a minimal footprint and maximum control.
• Flexible Backend Integration: Supports various AI agent backends like codex, claude, copilot, and grok on a per-role basis.

Typical Use Cases

SwarmForge is ideal for projects requiring:

• Automated Software Development: From rapid prototyping with the two-pack workflow to comprehensive development cycles with the six-pack, SwarmForge manages the entire lifecycle.
• Disciplined Code Refinement: The coder and cleaner roles in two-pack ensure code quality through TDD and automated reviews.
• Structured Specification & Implementation: The four-pack workflow offers a balanced approach, integrating Gherkin specifications with coding and architectural oversight.
• Complex Project Management: The six-pack workflow provides granular control over each quality gate, suitable for large-scale projects demanding rigorous testing and architectural integrity.
• Agent Team Coordination: Any scenario where multiple AI agents need to collaborate on a shared codebase and task, ensuring efficient and conflict-free progress.