cross-modal-review

You can get a second opinion by requesting a cross-modal review, which routes your work product to a different AI model for quality evaluation. The skill applies structured review-mode gating to determine when to invoke the review process before finalizing significant modifications.

Cross-modal AI review is a quality assurance gate that uses a different AI model to review your work, ensuring quality and agreement before finalizing. It applies to code reviews, security-sensitive changes, and other significant modifications.

You need the Codex CLI specifically for code reviews, as the skill offers a handoff to the Codex code-review skill for diff review. Access to different AI models is also required to perform the cross-modal review process.

If an AI model refuses to review your code, the skill silently switches to the next available model to complete the review. This refusal routing ensures the quality assurance process continues without interruption.

You should use a different AI model for code review before committing code changes, reviewing security-sensitive changes, or finalizing other significant modifications. This ensures quality and agreement across different models.