Coding Agent Orchestrator
Delegates complex development tasks and feature builds to specialized background AI agents like Codex and Claude Code.
The Coding Agent skill empowers users to offload intensive development work—such as building new features, performing large-scale refactors, or conducting batch PR reviews—to specialized AI agents running in the background. By utilizing pseudo-terminal (PTY) support and isolated working directories, it ensures interactive CLI agents function correctly while allowing the primary user to maintain their workflow. It excels at managing long-running tasks across multiple branches or repositories using git worktrees for safe, parallel execution.
Key Features
01Real-time session monitoring with log, poll, and kill capabilities
02Multi-agent support including Codex, Claude Code, Pi, and OpenCode
03PTY-enabled background process management for interactive terminal apps
04228,567 GitHub stars
05Isolated execution environments via workdir and git worktree integration
06Automated completion notifications via system event triggers
Use Cases
01Running parallel AI-driven code reviews across multiple open pull requests
02Building entire features or boilerplate apps from scratch in isolated directories
03Executing deep codebase refactors without blocking the main terminal session