Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Codex Delegation Skill

Delegates coding, code review, and plan review to the OpenAI Codex CLI from inside Claude Code, with disciplined sandbox levels.

Dev & CodingIntermediate★ 1,656⑂ 241AI score 9/10Last updated: Sep 27, 2026

What it does

  • Detects requests like "use codex", "let gpt implement this", or "give me a second-opinion review", composes the right codex exec command, runs it, and relays the result.
  • Picks the sandbox tier for you: read-only for analysis, --full-auto (workspace-write) for edits, and danger-full-access only with explicit per-task user consent.
  • Presents review findings by severity and never auto-applies fixes — it stops and asks which issues you actually want changed.
  • Handles long runs by backgrounding with tee /tmp/codex-*.log, plus the < /dev/null trick so non-TTY runs don't hang on stdin.
  • Ships reference docs: full CLI flag reference, prompting patterns, adversarial/plan review workflows, and a JSON findings schema.

Who it's for

  • Developers who want cross-model verification (Claude + GPT) instead of trusting a single model.
  • Anyone offloading large implementation or refactor jobs to a non-interactive agent.
  • Teams building a pre-PR adversarial review or pre-implementation plan-review ritual.

Examples

  1. "Have codex implement user authentication" → runs codex exec --full-auto "implement the user authentication feature" and summarizes files changed, tests run.
  2. "Get GPT to review my changes against main" → codex exec review --base main, findings ordered by severity with exact file:line, no edits applied.
  3. "Second-opinion review of this design before I build it" → uses the plan-review workflow and bundled JSON schema to emit structured risks and open questions.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/feiskyer/claude-code-settings/HEAD/skills/codex-skill/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/codex-skill folder from the GitHub repo feiskyer/claude-code-settings into my ~/.claude/skills/codex-skill/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/feiskyer/claude-code-settings.git /tmp/ccs && mkdir -p ~/.claude/skills && cp -r /tmp/ccs/skills/codex-skill ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal with Node.js or Homebrew available.
  2. Install the Codex CLI: npm i -g @openai/codex (or brew install codex).
  3. Verify with codex --version and complete the first-run OpenAI authentication.
  4. Clone the skill repo: git clone https://github.com/feiskyer/claude-code-settings.git /tmp/ccs
  5. Copy the skill: mkdir -p ~/.claude/skills && cp -r /tmp/ccs/skills/codex-skill ~/.claude/skills/
  6. Confirm the references/ and assets/ folders came along with it.
  7. Restart Claude Code and test with something like "use codex to analyze this repo's structure".
  8. For sensitive repositories, keep the read-only sandbox as the default and grant danger-full-access only when you truly need it.
View source on GitHub ↗License: MIT