codex-app-server-persisted

Manage a long-lived codex app-server process over stdio via JSON-RPC.

2|Updated Jan 30, 2026
One-click install
npx skills add https://github.com/tankygranny05/agent-box --skill codex-app-server-persisted
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-app-server-persisted
Source: https://github.com/tankygranny05/agent-box/tree/main/agent-box/seed/codex-skills/codex-app-server-persisted
Command: npx skills add https://github.com/tankygranny05/agent-box --skill codex-app-server-persisted

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the need for a consistently running Codex process to minimize latency in multi-turn interactions, avoiding the overhead of starting a new process for each turn.

Core Features & Use Cases

  • Persistent Process: Maintains a single codex app-server process alive across multiple turns.
  • Programmatic Control: Enables JSON-RPC communication over stdio for deterministic turn management.
  • Observability: Preserves sse_lines.jsonl for monitoring.
  • Use Case: Ideal for building chatbots or agents that require rapid responses by keeping the underlying Codex model warm and ready.

Quick Start

Use the codex-app-server-persisted skill to start a persistent Codex app server and begin a new thread for interaction.

Frequently Asked Questions about codex-app-server-persisted

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

FAQPage Schema
How do I keep a codex app-server process warm for low-latency multi-turn interactions?

To keep a codex app-server process warm for low-latency interactions, you can manage a long-lived process over stdio. This avoids the overhead of starting a new process for each turn, keeping the model ready.

How does JSON-RPC programmatic control work with a persistent Codex app-server?

JSON-RPC programmatic control with a persistent Codex app-server works by communicating over stdio. This enables deterministic turn management, thread management, and event streaming while maintaining the subprocess stdin.

Can I preserve sse_lines.jsonl observability when managing a persistent Codex app-server?

Yes, you can preserve sse_lines.jsonl observability when managing a persistent Codex app-server. The process maintains this file specifically for monitoring event streaming and multi-turn interactions.

What is the best way to manage a long-lived Codex app-server subprocess for chatbot agents?

The best way to manage a long-lived Codex app-server subprocess for chatbot agents is by spawning it as a subprocess and maintaining its stdin. This ensures rapid responses by keeping the underlying Codex model warm.

Do I need to spawn codex app-server as a subprocess to maintain persistent process threads?

Yes, you need to spawn codex app-server as a subprocess and maintain its stdin to facilitate programmatic control via JSON-RPC. This is required for thread management and low-latency multi-turn interactions.

Why does starting a new Codex process for each turn cause latency in multi-turn interactions?

Starting a new Codex process for each turn causes latency in multi-turn interactions due to process initialization overhead. Maintaining a single persistent process alive across multiple turns eliminates this delay.