takaaki-s

Jind AI — AI skill for Claude Code

AI community

Ten agent sessions, one screen.

How to install Jind AI

This entry records only its repository, not the path inside it, so there is no exact command to give. Open takaaki-s/jind-ai and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Jind AI does

Ten agent sessions, one screen. A tmux-based TUI for Claude Code, Codex and opencode — see which agent is waiting on you, give each its own git worktree, and resume after a reboot.

Alternatives in AI

Model Update — 增量更新财务模型 — 用新财报/新指引/修正假设重算 DCF 内在价值、Comps 隐含价、投资逻辑与目标价，输出 before→after delta 6.6k ★
Cwc Long Running Agents — Claude Code's built-in /goal command gives you a generator/evaluator loop out of the box: set a completion con 668 ★
Ntm — Named Tmux Manager: spawn, tile, and coordinate multiple AI coding agents (Claude, Codex, Gemini) across tmux 257 ★

README

**English** | [日本語](README.ja.md)

jind-ai

**Ten agent sessions. One screen.**

Never miss the agent that's waiting on you
A git worktree of its own for every session
Reboot the machine, pick up the conversation
Agents can drive agents

Runs on tmux, so your config carries over — and over SSH you come back to the same screen.

![Eight agent sessions in one list. The cursor moves down it and the pane below follows, describing each session without switching to it; the one waiting on a permission prompt is marked, and its prompt is open on the right.](https://github.com/takaaki-s/jind-ai/releases/download/v0.10.0/demo.gif)

**For AI agents.** jin's reference material for agents is not in this file — it is compiled into the binary, so it always matches the version installed. `jin docs list` names the topics, `jin docs show ` reads one, and neither needs the daemon. With only the repository in front of you, the same documents are [`internal/agentdocs/docs/`](internal/agentdocs/docs/); start with [orchestration](internal/agentdocs/docs/orchestration.md).

Installation

Download from GitHub Releases

Download the binary for your OS/architecture from the [Releases page](https://github.com/takaaki-s/jind-ai/releases).

# Example: Linux amd64
curl -Lo jind-ai.tar.gz https://github.com/takaaki-s/jind-ai/releases/latest/download/jind-ai_0.7.0_linux_amd64.tar.gz
tar xzf jind-ai.tar.gz
sudo mv jin /usr/local/bin/

Go install

go install github.com/takaaki-s/jind-ai/cmd/jin@latest

Build from source

git clone https://github.com/takaaki-s/jind-ai.git
cd jind-ai
make build    # Build to bin/jin
make install  # Install to $GOPATH/bin

What you can do

**Know which one needs you.** Status is reported by the agent itself — thinking, idle, or stopped waiting for your permission — rather than guessed by reading the screen.

**Keep agents out of each other's way.** `--worktree` cuts a git worktree and a br