run-codex

name	run-codex
description	Send a prompt to Codex (OpenAI) and get a response. Use for code review, refactoring, architecture questions, or any task where a second model's perspective is valuable. Synthesize Codex output with your own analysis.
allowed-tools	Bash, Read
argument-hint	<prompt>

Codex CLI

Send $ARGUMENTS as a prompt to Codex via codex exec. The entire argument string is the prompt — no mode parsing needed.

How to run

cd "$(git rev-parse --show-toplevel)"
OUTFILE=$(mktemp /tmp/codex_result.XXXXXX)
/opt/homebrew/bin/codex exec "$ARGUMENTS" \
  -m gpt-5.3-codex \
  -c model_reasoning_effort="high" \
  --ephemeral --full-auto \
  -o "$OUTFILE" 2>&1

Read $OUTFILE for the output. If the exit code is non-zero, read $OUTFILE for the error.

After running

1. Present Codex's output under a Codex heading
2. Add your own take under a Claude heading
3. Note agreements, disagreements (with resolution), and blind spots
4. List concrete next steps