Claude Code Orchestrator for Openclaw

A tmux-based orchestration tool for managing long-running Claude Code development tasks with high visibility and automated callbacks.

yaxuan42
v0.1.0
Feb 15, 2026
0
1.9k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install claude-code-legacy

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install claude-code-legacy using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Claude Code Orchestrator?

The Claude Code Orchestrator is a robust management layer designed to handle complex, long-duration coding tasks using the Claude CLI. By leveraging tmux sessions, it ensures that AI-driven development work does not hang or fail silently. This skill is a vital part of the Openclaw Skills ecosystem, providing developers with a stable environment to monitor progress, SSH into active sessions, and receive automated notifications upon task completion.

It bridges the gap between raw CLI execution and observable agentic workflows by providing a structured lifecycle for every coding job. Whether you are performing a massive refactor or a design system update, this orchestrator keeps the process transparent and recoverable.

Claude Code Orchestrator Use Cases

  • Running large-scale refactoring tasks that require hours of Claude Code execution.
  • Monitoring AI coding progress in real-time via SSH or tmux attach.
  • Automating notifications and summaries for completed development cycles.
  • Preventing silent failures in headless AI agent environments.
  • Managing multiple concurrent AI-led coding sessions with structured status reporting.

How Claude Code Orchestrator Works

  1. The system creates a dedicated prompt file to prevent shell quoting errors during task initialization.
  2. A specialized tmux session is launched to house the interactive Claude CLI environment.
  3. Claude is initialized with specific flags to skip manual permission prompts and begin processing the task.
  4. The prompt is injected into the session, including a mandatory callback script that triggers once work is finished.
  5. Users can monitor the session via tmux sockets or receive a structured JSON summary of all active tasks using Openclaw Skills utilities.
  6. Upon completion, a mandatory review loop analyzes the diffs, linting results, and risk summaries before notifying the user.

Claude Code Orchestrator Setup

To get started with this skill within the Openclaw Skills framework, ensure you have tmux installed and then configure your base directory. Use the following commands to initialize and monitor tasks:

# Start a new orchestrated coding task
bash {baseDir}/scripts/start-tmux-task.sh \
  --label "feature-update" \
  --workdir "/path/to/project" \
  --prompt-file "/path/to/prompt.md"

# Monitor an active session
bash {baseDir}/scripts/monitor-tmux-task.sh --attach --session <session_name>

Claude Code Orchestrator Data Schema & Taxonomy

The skill maintains a structured data hierarchy to track task evolution and reporting, often processed through Openclaw Skills scripts.

Property Description
label A unique human-readable identifier for the coding task.
status Current state: running, stuck, idle, dead, or likely_done.
reportJsonPath Location of the generated JSON completion reports in /tmp/.
lastLines A buffer of the most recent terminal output from the tmux pane.
sessionAlive Boolean flag indicating if the underlying tmux session is still active.

Claude Code Orchestrator Advanced Features

  • Automated status recovery to detect stuck or idle Claude sessions without consuming extra tokens.
  • Structured JSON output for all task listings, enabling integration with butler-style summary agents.
  • Robust prompt submission logic with ready-check and multi-enter retry mechanisms to avoid submission failures.
  • Mandatory deep-read completion loops that synthesize transcript history with current user chat context.
  • Remote host support via SSH aliases for managing tmux sessions on external development servers.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*