codex-invoke

Spawn roster agent personas in a cmux pane with OpenAI Codex routing.

1|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/firefly-events/plugin-hive --skill codex-invoke
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-invoke
Source: https://github.com/firefly-events/plugin-hive/tree/main/skills/hive/skills/codex-invoke
Command: npx skills add https://github.com/firefly-events/plugin-hive --skill codex-invoke

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Spawn roster agent personas in a Codex-enabled pane to access OpenAI Codex capabilities while keeping the Claude path decoupled and ensuring consistent memory and prompts.

Core Features & Use Cases

  • Side-by-side Codex execution: run a roster persona in a visible cmux pane with Codex routing to reduce cost and model bias.
  • Controlled pre-flight and routing: enforces platform constraints (macOS) and persona compatibility rules to decide when Codex is used or bypassed.
  • Pane lifecycle management: supports one-shot and persistent pane modes with transcript and meta capture for auditing and debugging.

Quick Start

Open a Codex pane and execute the selected roster persona with the given inputs to produce a transcript.

Frequently Asked Questions about codex-invoke

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

FAQPage Schema
How do I run Codex agents in a visible terminal pane alongside Claude?

You can invoke Codex agents in a visible side-by-side cmux pane using codex-invoke, which spawns a roster agent persona and routes execution through OpenAI Codex while keeping the Claude path decoupled.

Can I use Codex routing on Linux or Windows to reduce model bias?

No, Codex routing through this method is governed by platform constraints and currently applies only when the target backend is Codex on macOS, bypassing Codex on unsupported operating systems.

How does pane lifecycle management work for Codex-backed roster agents?

Pane lifecycle management supports both one-shot and persistent pane modes, handling idle-timeout automatically while capturing transcript and metadata for auditing and debugging after Codex execution.

What pre-flight checks are enforced before routing execution to OpenAI Codex?

Controlled pre-flight checks enforce macOS platform constraints and verify roster persona compatibility rules to determine whether Codex routing is used or bypassed before system prompt assembly.

Why route roster agent personas through Codex instead of using Claude directly?

Routing roster agent personas through Codex reduces cost and model bias while maintaining consistent memory and prompts, ensuring the Claude path remains decoupled from OpenAI Codex execution.