codex-dispatch

Dispatch tasks to OpenAI Codex CLI and GPT models for review and implementation.

Updated May 28, 2026
One-click install
npx skills add https://github.com/SensLiao/Claude-code-setting --skill codex-dispatch-sensliao
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-dispatch
Source: https://github.com/SensLiao/Claude-code-setting/tree/main/skills/codex-dispatch
Command: npx skills add https://github.com/SensLiao/Claude-code-setting --skill codex-dispatch-sensliao

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires codex@openai-codex, and includes scripts (resource) components.

What problem does it solve?

This Skill provides a decision-making and discipline layer for effectively utilizing Codex (OpenAI Codex CLI, GPT series models) as a secondary model family for delegated execution and cross-model review.

Core Features & Use Cases

  • Decision-Making: Determines when and how to use Codex, with a decision tree for choosing the right commands and enforcing discipline.
  • Routing: Routes natural language intents to the appropriate /codex:* commands.
  • Discipline: Enforces harness-specific discipline with rules like when-to-delegate decision trees, work-order templates, and model-version locks.

Quick Start

Dispatch a task to Codex using the command: /codex:rescue "My task description".

Frequently Asked Questions about codex-dispatch

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

FAQPage Schema
How do I route natural language tasks to OpenAI Codex CLI for delegated execution?

You can route natural language intents to Codex by dispatching tasks through specific commands, which triggers a decision tree that chooses the right execution path and enforces model discipline for the delegated work.

What is cross-model review and when do I need it for code review and bug fixing?

Cross-model review uses Codex as a secondary model family to evaluate tasks like code review and bug fixing. You need it when you want a distinct model to validate implementations and enforce discipline across your workflow.

How to enforce model-version locks and discipline when dispatching code execution tasks?

You enforce model-version locks by applying a discipline layer that utilizes work-order templates and decision trees. This ensures the Codex CLI executes tasks strictly within the defined harness rules and version constraints.

Do I need the codex plugin to use cross-model review and routing features?

Yes, you need the codex plugin. The decision-making and routing features require the codex plugin to facilitate task execution, coordinate cross-model review, and dispatch commands effectively.

Can I use Codex dispatch for both code implementation and bug fixing?

Yes, you can use it for both. The dispatch layer routes natural language intents to appropriate commands, coordinating decision-making and execution for code implementation, bug fixing, and code review tasks.

What's the best way to delegate a task to Codex using a rescue command?

The best way to delegate a task is to use the rescue command with a natural language string. This triggers the routing layer to interpret the intent and apply the correct execution discipline for the task.