codex-exec

Execute OpenAI Codex CLI prompts in ephemeral environments with JSON output.

1|Updated May 21, 2026
One-click install
npx skills add https://github.com/hiddink-ai/hiddink-harness --skill codex-exec-hiddink-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-exec
Source: https://github.com/hiddink-ai/hiddink-harness/tree/main/templates/skills/codex-exec
Command: npx skills add https://github.com/hiddink-ai/hiddink-harness --skill codex-exec-hiddink-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This skill resolves the friction of coordinating between Claude and OpenAI Codex by providing a standardized, non-interactive execution bridge that handles authentication, environment validation, and structured output parsing.

Core Features & Use Cases

  • Hybrid Execution: Enables Claude to delegate complex code generation or research tasks to the Codex CLI while maintaining a unified conversation context.
  • Structured Results: Supports JSON Lines output for programmatic parsing of assistant messages and execution metadata.
  • Safety & Control: Includes built-in timeout management, ephemeral session handling, and optional full-auto approval modes for high-effort tasks.

Quick Start

Use the codex-exec skill to generate a new project README file with high reasoning effort and return the result in JSON format.

Frequently Asked Questions about codex-exec

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

FAQPage Schema
How do I run Codex CLI prompts non-interactively for automated workflows?

You can execute Codex CLI prompts non-interactively by delegating them to an ephemeral environment that validates binary availability, checks authentication status, and processes structured JSON output for reliable automated task completion.

Can I use Claude to delegate code generation tasks to OpenAI Codex?

Yes, hybrid AI execution bridges Claude and OpenAI Codex by enabling Claude to delegate complex code generation or research tasks to the Codex CLI while maintaining a unified conversation context.

What is structured JSON Lines output when executing CLI code generation tasks?

Structured JSON Lines output provides programmatic parsing of assistant messages and execution metadata, allowing downstream systems to consume the results of delegated code generation or research tasks reliably.

Does Codex CLI execution support configurable reasoning effort for research tasks?

Yes, Codex CLI execution supports configurable reasoning effort levels, allowing you to adjust the depth of processing for research or code generation tasks within the non-interactive execution environment.

How do I validate OpenAI Codex authentication before running automated prompts?

The execution bridge validates binary availability and authentication status before delegating any prompts to the Codex CLI, ensuring that only authenticated and properly configured sessions are executed.

What are the limitations of using ephemeral environments for CLI orchestration?

Ephemeral environments handle session isolation and timeout management, but require valid authentication and available binaries before execution; tasks failing these prerequisite checks will not be delegated to the Codex CLI.