skill-codex
A Claude Code skill that delegates tasks to OpenAI's Codex CLI for code analysis, review and edits
- 1,478 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Runs Codex CLI, which can change files and execute commands
- In full-access mode Codex reaches the network
Install
- Works in
Manual install
/plugin marketplace add skills-directory/skill-codex
/plugin install skill-codex@skill-codexCodex CLI must be installed and available on PATH.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Runs Codex CLI, which can change files and execute commands
- In full-access mode Codex reaches the network
This is third-party code. Review the repository files before installing.
What it does
The skill teaches Claude Code to call OpenAI's Codex CLI. On request the agent assembles a codex exec invocation: it picks the model and reasoning effort, sets the sandbox mode from read-only to full access, runs the command and summarizes the result. It supports continuing a previous session with resume and by default hides service output so it does not bloat the context. It installs as a Claude Code plugin or by copying the skill folder into the skills directory.
Who it is for. For developers in Claude Code who want to bring Codex CLI into tasks as a second model.
Good fit when
- You need Codex to analyze a repository or propose edits
- You want a review or refactor from a second model
- You need to resume a previous Codex session on the code
Not a fit when
- Codex CLI is not installed or there is no access to OpenAI models
- One agent is enough without a second model
Example request
Use codex to analyze this repository and suggest improvementsLimitations
You need Codex CLI installed with valid access to OpenAI models. OpenAI models are not reachable from Russia without a VPN. In write and full-access modes Codex changes files and reaches the network, so run it deliberately.
How to disable. Remove the plugin with /plugin, or delete the codex folder from the ~/.claude/skills/ directory.
FAQ
What is needed besides the skill?
Codex CLI on PATH with configured access to OpenAI models.
Can Codex change code?
Yes, in write or full-access mode. By default it uses read-only mode.
Discussion
No questions or reviews yet.
Related
SkillsA Claude Code skill that drives Codex over JSON-RPC: tasks, code reviews, parallel research and messaging between two agents.
lencx skills
lencx-skills
Two skills for Claude Code and Codex: keel works through architecture questions, coding-protocol keeps code edits inside the task
A Claude Code plugin: a /codex skill and a hook that sends every finished plan to Codex CLI for review before you approve it.
A skill for second-agent review of plans or code: Claude Code starts rounds, Codex reviews, and they exchange files
OpenAI's official plugin: Codex reviews of your changes and task delegation to Codex from inside a Claude Code session
61 small skills for Codex across 13 areas: codebase orientation, tests, migrations, security, CI and pull request review
README in short
The README presents the repository as a Claude Code plugin with a marketplace and explains how the skill calls Codex CLI via codex exec and session resume. Before use you install Codex CLI and check it with codex --version. Installation is either as a plugin through the marketplace or by manually copying the skill folder. The skill lets you choose the model, reasoning effort and sandbox mode, hides service output by default, and advises treating Codex answers critically.
README badge
Original SKILL.md
--- name: codex description: Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing --- # Codex Skill Guide ## Running a Task 1. For a new session (resumes inherit the prior model/effort), ask the user which model AND which reasoning effort to use, in a single prompt with two questions. When the user expresses no preference, use the documented defaults. 2. Select the sandbox mode required for the task; default to --sandbox read-only unless edits or network access are necessary. 3. Assemble the command with the appropriate options: model, reasoning effort, sandbox, and the prompt as the final positional argument. 4. Always use --skip-git-repo-check. 5. When continuing a previous session, resume via stdin without configuration flags unless the user asks for them.