coding-agent

Delegate coding tasks to Codex, Claude Code, or Pi agents via background processes.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/kulharir7/karya --skill coding-agent-kulharir7
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: coding-agent
Source: https://github.com/kulharir7/karya/tree/main/workspace/plugins/coding-agent
Command: npx skills add https://github.com/kulharir7/karya --skill coding-agent-kulharir7

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The coding workflow can be slowed by manual, repetitive setup when delegating coding tasks to AI agents; this skill provides a controlled way to offload coding work to Codex, Claude Code, Pi, or other agents via a background process.

Core Features & Use Cases

  • Orchestrates background AI coding sessions with PTY-enabled terminals.
  • Supports multiple agent backends (Codex, Claude Code, Pi, OpenCode) with per-task configuration.
  • Use Case: building features, reviewing PRs in isolated temp dirs, refactoring large codebases, or iterative exploration of code.

Quick Start

Create a temporary git repository, start a coding agent in background, and issue your coding task.

Frequently Asked Questions about coding-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I delegate coding tasks to an AI agent in the background?

You can delegate coding tasks to AI agents in the background using this Skill, which orchestrates background sessions with agents like Codex, Claude Code, Pi, or OpenCode to handle feature building, PR reviews, and refactoring.

Can I use Claude Code for background coding tasks without a PTY terminal?

Claude Code supports background coding without a PTY by using the --print --permission-mode bypassPermissions flags, while Codex, Pi, and OpenCode require pty:true terminal configurations.

What is the best way to review PRs using an AI coding agent?

Reviewing PRs with an AI agent is best handled by checking them out into temporary directories and delegating the review to a background process, which this Skill orchestrates to ensure isolated and safe code analysis.

Does this background AI agent support multiple coding backends for refactoring large codebases?

This Skill supports multiple coding backends for refactoring large codebases, including Codex, Claude Code, Pi, and OpenCode, allowing per-task configuration to match the specific agent to your refactoring needs.

Do I need a PTY-enabled terminal to run OpenCode in the background?

You need a PTY-enabled terminal to run OpenCode in the background because it requires pty:true, whereas Claude Code bypasses this requirement by using the --print --permission-mode bypassPermissions flags.