adversarial-review

Cross-model code review routes code changes to a different AI model for independent validation. This enforces model-divergent review to surface blind spots by ensuring the AI reviewer differs from the code author model, returning structured findings for PR workflows.

Use adversarial AI review during critical pull requests and security-sensitive code changes. Cross-model validation surfaces blind spots that a single AI model might miss, providing independent findings for safer code integration.

Submit a code change and request an independent cross-model review. The Skill routes the changes to a different AI model than the code author, enforcing model-divergent review and returning structured findings for security-sensitive code.

Yes, cross-model validation returns structured findings designed for integration with PR workflows. This allows you to compare external model findings with your internal review context during critical pull requests.

No specific dependencies or components are required. The Skill operates independently to route code changes to a different AI model for review, enforcing model-divergent validation without needing external modules or setup.