bmad-review-adversarial-general

An adversarial review systematically tests task completions against predefined criteria to validate robustness and correctness. It ensures quality assurance by following detailed procedural steps to identify potential flaws in outputs.

To execute an adversarial review, provide a specified XML task file containing detailed instructions. The review process reads and follows these predefined procedural steps to systematically validate the task completion.

Yes, this adversarial review is applicable to quality assurance and validation workflows. It systematically checks the robustness and correctness of AI-generated outputs or general task completions against specified criteria.

Yes, you need a specified XML task file to perform the review. The Skill requires adherence to the detailed procedural steps and validation criteria outlined within that task file to function correctly.

The best way to validate AI outputs is using an automated adversarial review process. By defining criteria in a task file, you can systematically check task completions for robustness and adherence to expected standards.