cmuxlayer
Active·★ 30·Apache-2.0·Updated 2026-10-22
cmuxLayer addresses the challenge of AI agents operating in isolated terminal environments by providing 35 Model Context Protocol (MCP) tools. It enables AI agents to programmatically control terminal workspaces, manage agent lifecycles, and facilitate multi-agent orchestration for enhanced collaboration.
#AI Agent Tools#Terminal Automation#Workspace Management#Multi-Agent Orchestration#Developer Tooling#Model Context Protocol
01
Features
01Terminal Workspace Control: Provides programmatic control over terminal workspaces, including creating splits, new tabs, sending input, and reading screen output.
02Multi-Agent Orchestration: Enables spawning, routing tasks to, and waiting for agents, facilitating complex multi-agent workflows.
03Agent Metadata Parsing: Auto-detects and parses agent metadata (status, model, tokens, context %) from terminal output for supported agents.
04Integrated with MCP: Exposes functionality through 35 Model Context Protocol (MCP) tools for seamless integration with AI agent frameworks.
02
Compatibility
cmux
Core Dependency
Verified via docs
Claude Code
AI Agent
Verified via docs
Codex CLI / T3 Code
AI Agent / IDE
Verified via docs
Gemini CLI
AI Agent
Verified via docs
Cursor
AI Agent / IDE
Verified via docs
VS Code
IDE
Verified via docs
03
Quick start
1
$ brew install cmux
2
$ brew install etanhey/layers/cmuxlayer
04
Use cases
↳Automated Test Execution: Split a pane to run a test suite, allowing an AI agent to monitor test progress.
↳Code Refactoring by Agents: Spawn a new agent in a dedicated pane to refactor specific code files.
↳Build Status Monitoring: Read the screen of a specific terminal surface to check if a build has passed.
↳Multi-Agent Task Synchronization: Wait for all agents to complete their tasks, then collect and process their outputs.
05
Alternatives
claudexor★ 516
Local-first multi-harness control plane for Claude Code, Codex, Cursor, and OpenCode: quota-aware rotation across multiple Claude/Codex subscriptions, shared thread context, and cross-model review.
polter★ 20
One Claude Code session supervises the agents in your other terminal tabs: which is stuck, which is done, which is still running. A fork of Ghostty.
entwurf★ 29
Garden-citizen dispatch for pi, Claude Code, Copilot CLI, OMP, Antigravity, and Codex; visible tmux sibling seats for Pi/Claude/Copilot/OMP/Codex, preserving each harness's auth, tools, and transcript; plus Claude/Cortex ACP.
omk★ 148
Evidence-gated runner for Codex, Claude Code, OpenCode, and local coding agents. Routes tasks into scoped DAG lanes with replayable artifacts.
open-multi-agent-kit★ 148
Provider-neutral multi-agent control plane for coding agents. Route runtimes, scope MCP tools, run DAG workers, verify evidence, and replay agent runs from the omk CLI.
itasca-mcp★ 219
MCP server connecting AI agents to ITASCA engines (PFC, FLAC, 3DEC, MPoint, MassFlow) — run numerical simulations through natural conversation
agentic-kit★ 61
One command to install, heal & prove ruflo (claude-flow) + agentic-qe — plus Claude + Codex dual-host with per-activity model routing. Thin, reversible, cross-platform.
Agentflow★ 27
Production-grade framework for building multi-agent AI systems. Graph-based orchestration, LLM-agnostic (OpenAI, Google GenAI, Anthropic), 3-layer memory (Redis cache + Postgres + vector store), live agents, parallel tool execution, and native MCP. Ships a full ecosystem: backend, REST API + CLI, TypeScript SDK, and React playground
Comments
Log in to leave a comment
No comments yet. Be the first!