coding-agent

Run interactive coding agents in PTY sessions with process controls.

Updated Mar 21, 2026
One-click install
npx skills add https://github.com/Polealpha/XINNIAN_PI --skill coding-agent-polealpha
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: coding-agent
Source: https://github.com/Polealpha/XINNIAN_PI/tree/main/app%20windows/vendor/openclaw-runtime/skills/coding-agent
Command: npx skills add https://github.com/Polealpha/XINNIAN_PI --skill coding-agent-polealpha

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the problem of getting interactive coding agents (Codex, Claude Code, OpenCode, or Pi Coding Agent) to run reliably via automation without broken output or hangs.

Core Features & Use Cases

  • Interactive PTY execution: Ensures interactive CLI agents get a pseudo-terminal so they behave correctly.
  • One-shot and background sessions: Supports quick one-off prompts and longer-running background workflows with session tracking.
  • Process-style control loop: Enables listing, polling, logging, and safely interacting with running agent sessions (including input submission and termination).

Quick Start

Start a one-shot Codex chat in a real project directory (with a git repo) using PTY so the agent runs correctly end-to-end.

Frequently Asked Questions about coding-agent

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

FAQPage Schema
Why do interactive coding agents like Codex hang or break when run via automation?

Interactive coding agents hang in automation without a pseudo-terminal (PTY). Allocating a PTY ensures interactive CLI agents execute correctly end-to-end without broken output or unexpected stalls.

How do I run a one-shot coding agent prompt in a real project directory?

To run a one-shot coding agent prompt, execute the agent in your target git repository directory with PTY allocation enabled. This allows the agent to process the single prompt and produce programmatic code changes safely.

Can I manage long-running coding agent sessions in the background?

Yes, you can manage background sessions for long-running coding agents. The process control loop supports listing, polling, logging, and safely interacting with running agent sessions, including input submission and termination.

Does this PTY control method work with full-auto builds and iterative development tasks?

Yes, PTY control applies to one-shot coding prompts, full-auto builds, and background session workflows for iterative development tasks. It ensures reliable interactive CLI agent execution across various automated coding workflows.

What process actions are available for controlling background agent sessions?

Available process actions for controlling background agent sessions include log, poll, write/submit, and kill. These actions enable tracking, interacting with, and safely terminating running coding agent sessions.