Codex Review

Optional cross-vendor executor. Requires the OpenAI codex CLI. If command -v codex fails, this skill isn't available on this machine — say so and fall back to a native review. The model-orchestration doctrine (references/model-orchestration.md) routes here only when the CLI is present.

Use Codex as an independent reviewer when the user wants a second-pass review or when a change is broad enough that another agent's perspective is useful.

Prefer Claude's normal review process for small local checks. Do not delegate review just to avoid reading the code yourself. Treat Codex's output as evidence, not authority.

Workflow

1. Identify the review target: uncommitted changes, base branch, commit SHA, PR checkout, or specific files.
2. Create a temporary artifact directory for the Codex report.
3. Run codex review with a focused review prompt.
4. Read Codex's report and verify important claims against the code before presenting them.

Use one of these command shapes: