claude-colony
Multi-agent Claude Code you can actually control. Spawn specialist teams in tmux, watch them work side-by-side, steer when needed. Built on native Claude Code agents. No black boxes.
claude mcp add --transport stdio makingjamie-claude-colony npx -y claude-colony
How to use
Claude Colony is a tmux-based multi-agent orchestration system that coordinates multiple Claude Code instances to work together on software development tasks. It leverages Claude Code's native agent system, where agents live as markdown prompts in the .claude/agents/ directory and are orchestrated into teams through colony presets. The colony provides a visual, file-based communication workflow via a shared message broker and tmux panes, allowing agents to collaborate in real time, assign roles, and coordinate complex tasks across frontend, backend, API, DevOps, and other specialties. You can start with built-in presets like fullstack, frontend-team, or api-team, or customize your own team by creating agents in your repository and building a corresponding preset. The system emphasizes transparency and parallel work, enabling agents to spawn sub-agents, use Claude Code tools, and communicate through a simple @mention protocol in a conversational, developer-friendly environment.
How to install
Prerequisites:
- Node.js 18+ and npm
- tmux installed (e.g., brew install tmux on macOS or apt install tmux on Linux)
- Claude Code installed and authenticated
Installation steps:
-
Install Claude Colony globally (preferred method): npm install -g claude-colony
-
Verify installation or run via npx: npx claude-colony --help
-
If you want to build from source (optional): git clone https://github.com/jamiemurphy/claude-colony.git cd claude-colony npm install npm run build npm link
-
Start using Claude Colony with a preset: colony start --preset fullstack
Prerequisites recap:
- Ensure Node.js and npm are up to date
- Ensure tmux is installed and available in your PATH
- Have Claude Code installed and authenticated so agents can operate
Additional notes
Tips and notes:
- The colony relies on a file-based message broker and @mention style communication; keep your .claude/agents/ and colony presets under version control for reproducibility.
- Be mindful of token usage with Claude Code; large teams or long-running sessions can consume significant tokens.
- Start with a preset that matches your project structure, then customize agents in .claude/agents/ to tailor behavior to your needs.
- If you modify agents or presets, run colony build or colony init as needed to refresh the configuration.
- Common issues often relate to tmux pane management or missing Claude Code authentication; verify that Claude Code commands run correctly in your shell before starting the colony.
Related MCP Servers
dify
Production-ready platform for agentic workflow development.
ruflo
🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, distributed swarm intelligence, RAG integration, and native Claude Code / Codex Integration
solace-agent-mesh
An event-driven framework designed to build and orchestrate multi-agent AI systems. It enables seamless integration of AI agents with real-world data sources and systems, facilitating complex, multi-step workflows.
pilot-shell
The professional development environment for Claude Code. Claude Code is powerful. Pilot Shell makes it reliable. Start a task, grab a coffee, come back to production-grade code. Tests enforced. Context preserved. Quality automated.
agentic-radar
A security scanner for your LLM agentic workflows
RAGLight
RAGLight is a modular framework for Retrieval-Augmented Generation (RAG). It makes it easy to plug in different LLMs, embeddings, and vector stores, and now includes seamless MCP integration to connect external tools and data sources.