codex-subagents

Delegate bounded, independent tasks to Codex subagents for parallel execution.

37|2|Updated Mar 25, 2026
One-click install
npx skills add https://github.com/jpcaparas/skills --skill codex-subagents-jpcaparas
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-subagents
Source: https://github.com/jpcaparas/skills/tree/main/skills/codex-subagents
Command: npx skills add https://github.com/jpcaparas/skills --skill codex-subagents-jpcaparas

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill helps streamline and optimize workflows in Codex CLI/App by allowing users to delegate tasks to subagents, thereby improving efficiency and productivity.

Core Features & Use Cases

  • Explicit Authorization: Users can explicitly request subagents for specific tasks.
  • Bounded Delegation: Subagents are spawned for bounded, independent tasks that can run in parallel without blocking the main agent.
  • Write Ownership: Subagents are assigned specific write ownership to prevent conflicts and ensure proper integration.
  • Model-Neutral: Subagents are not tied to a specific model, allowing for flexibility and adaptability.
  • Use Case: When reviewing a codebase, users can spawn subagents to perform parallel reviews, research, and testing, while the main thread focuses on synthesis and verification.

Quick Start

To start using subagents in Codex CLI/App, install the codex-subagents skill and then use the command: npx skills add jpcaparas/skills --skill codex-subagents followed by the task you want to delegate.

Frequently Asked Questions about codex-subagents

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

FAQPage Schema
How do I delegate parallel tasks to subagents in Codex CLI?

To delegate parallel tasks in Codex CLI, you spawn subagents for bounded, independent operations that run without blocking the main thread. The main agent then focuses on local verification and synthesizing the results from these parallel workflows.

What is bounded delegation for parallel agents in code review?

Bounded delegation assigns specific write ownership to parallel agents during code review to prevent file conflicts. This ensures subagents can safely execute independent research, testing, and reviews simultaneously before the main thread integrates the changes.

Does the Codex subagents workflow require a specific AI model?

The Codex subagents workflow is model-neutral, meaning subagents are not tied to a specific AI model. This allows flexible, model-agnostic execution for delegated tasks within the Codex CLI or App environment.

When should I use parallel agents instead of a single agent in Codex?

Use parallel agents in Codex when you have independent, bounded tasks that can run simultaneously without blocking the main agent. This approach maximizes efficiency during complex operations like codebase reviews, where parallel research and testing are required.

Can I manage multiple subagents with write ownership in Codex App?

You can manage multiple subagents in Codex App by assigning them explicit write ownership for specific tasks. This bounded delegation prevents integration conflicts and allows the main thread to safely verify and synthesize their parallel outputs.

How to prevent conflicts when running parallel agents in Codex CLI?

To prevent conflicts when running parallel agents in Codex CLI, assign bounded write ownership to each subagent. This ensures independent tasks do not overlap, allowing the main agent to verify and synthesize the final outputs locally.