shuxueshuxue

shuxueshuxue / ccm-orchestra

Public

在 tmux 中管理交互式 Claude Code agent，并在 kitty 中协调不同 code agent 的可见协作与接力

18
2
69% credibility
Found Apr 07, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A coordination system for running persistent interactive AI coding sessions and enabling visible collaboration between multiple AI agents.

How It Works

1
🔍 Discover AI Coding Helpers

You find a handy tool that lets AI assistants stick around to help with your coding projects whenever you need them.

2
🛠️ Get Ready

You gather a few basic everyday tools on your computer to make everything work smoothly.

3
🚀 Start Your AI Teammate

With one simple command, you bring your first AI coding buddy to life right in your project folder.

4
💬 Assign a Task

You send a message to your AI buddy, like 'review this code and suggest improvements'.

5
⏳ Wait and Read Back

You check back after a short wait to see the thoughtful suggestions your AI buddy prepared.

6
🔄 Keep It Going

You reuse the same buddy for more tasks, or share messages with other helpers for teamwork.

✅ Smarter Coding Life

Your projects improve faster with reliable, always-ready AI help that feels like a real teammate.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ccm-orchestra?

CCM Orchestra runs persistent interactive Claude Code agents in tmux panes, letting you start named sessions like `ccm start frontend-agent`, send prompts with `ccm send`, and poll transcripts via `ccm read --wait-seconds 30`—all scoped to your working directory for repo isolation. It pairs this with kitty for visible collaboration, using commands like `ccm relay` to hand off tasks between tabs and `ccm tabs` to list peers with their git branches. Built in Python, it keeps agents reusable across sessions without relying on non-interactive modes that risk API limits.

Why is it gaining traction?

Unlike basic tmux claude code wrappers or github tmux plugins, it enables tmux agent team claude code workflows with push-based kitty relays and wechat-style messaging (`ccm wechat-send kitty:mycel`), making agent handoffs observable and mobile-friendly. The heartbeat tool pings supervision tabs to prevent silent deaths, and smoke tests verify end-to-end reliability. Devs dig the operational focus: doctor checks flag claude path mismatches post-upgrades, ensuring tmux claude code agents stay alive.

Who should use this?

Backend engineers coordinating Claude for code reviews across repos, frontend teams treating agents as write-enabled collaborators for layout critiques, or solo devs running tmux claude windows with kitty for multi-agent relays. Ideal for those scripting claude code setup with shift-enter flows or needing tmux claude code scroll without manual attaches.

Verdict

Try it if you're deep in Claude Code and tmux/kitty—solid docs, tests, and bilingual README make onboarding fast, despite 18 stars signaling early maturity. The 0.699999988079071% credibility score reflects its niche freshness, but `ccm-smoke` proves it works out of the box for tmux claude code github fans.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.