A background-first utility for executing and managing AI coding agents like Codex, Claude Code, and Pi for automated software development.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install coding-agentem9ak
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install coding-agentem9ak 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 Coding Agent?
The Coding Agent skill is designed for high-throughput development workflows, allowing users to run powerful AI agents like Codex CLI, Claude Code, and Pi Coding Agent as background processes. By utilizing Openclaw Skills, developers can delegate complex tasks such as feature building, refactoring, and automated testing to these agents without tying up their primary terminal session.
This skill emphasizes a workdir + background pattern, which ensures that agents operate within a focused directory context. This prevents the AI from accessing unrelated files and allows for parallel execution of multiple tasks, such as reviewing several pull requests simultaneously or fixing a batch of bugs across different branches.
Coding Agent Use Cases
- Running high-speed code generation tasks using Codex with the yolo or full-auto flags.
- Conducting parallel pull request reviews by deploying a fleet of background agents across different PR refs.
- Automating bug fixes across multiple issues using git worktrees to maintain isolation.
- Executing non-interactive one-shot prompts for codebase summarization or documentation updates.
How Coding Agent Works
- The user defines a specific working directory to constrain the agent's context and prevent it from wandering into unrelated files.
- A coding agent command is launched using the background mode, which returns a unique sessionId.
- The process is monitored asynchronously using log or poll actions to track the agent's progress.
- If the agent requires user feedback, the process action write command is used to send input to the session.
- Completed tasks are reviewed, and the session is terminated once the implementation or review is finalized.
Coding Agent Setup
Before using this skill, ensure that OpenClawCLI is installed on your system. You must also have one or more supported coding agents installed locally.
# Install Pi Coding Agent
npm install -g @mariozechner/pi-coding-agent
# Ensure your preferred agent (claude, codex, or opencode) is in your PATH
# Example of starting a background task:
bash workdir:~/my-project background:true command:"codex --yolo 'Add unit tests for the auth module'"
Coding Agent Data Schema & Taxonomy
The skill manages operations through process-based metadata and directory scoping to ensure safety and organization within Openclaw Skills.
| Attribute | Description |
|---|---|
| sessionId | A unique identifier for each background coding process. |
| workdir | The filesystem path where the agent is localized. |
| command | The specific CLI string passed to the underlying AI agent. |
| background | Boolean flag indicating if the task should run as a non-blocking process. |
Coding Agent Advanced Features
- Parallel PR Army: Ability to fetch multiple git refs and deploy individual agents to review every open PR simultaneously.
- Git Worktree Integration: Supports advanced multi-issue fixing workflows by isolating agents in separate git worktrees.
- Tmux Persistent Sessions: Provides TTY-capable environments for interactive coding agents that require a real terminal for history capture.
- Safety Sandboxing: Strict rules regarding directory access to prevent agents from interfering with the main project environment or sensitive system files.
SKILL.md
Loading
Related Openclaw Skills
Coding Agent
A background-first skill to programmatically control AI coding agents like Codex, Claude Code, and Pi for automated development workflows.
Coding Agent
Execute and manage top-tier AI coding agents as background processes for autonomous, parallel software development.
Coding Agent (Background-First)
A powerful skill for running AI coding agents like Codex and Claude Code in background processes for automated, parallel development tasks.
Clawdbot Logs
A comprehensive diagnostic suite for monitoring performance, analyzing errors, and tracking session statistics for Clawdbot.
Coding Agent (background-first)
A high-performance skill for executing AI coding agents in isolated background processes for automated software engineering and parallel code reviews.
Deep Research Agent
A sophisticated AI agent that decomposes complex inquiries into structured research plans for deep analytical insights.
Featured*
