red-team-review

Yes, it orchestrates multi-channel adversarial reviewers, including human reviewers, browser-based tools, and CLI environments. It bundles the research context and dispatches structured prompts to these isolated reviewers, maintaining iteration history until an Approved verdict is reached.

Manifest-based bundling aggregates your research context, design documents, and deterministic prompts into a structured review packet. This ensures isolated iterations and explicit completion signals, allowing adversarial reviewers to critique the architecture or security design effectively across multiple rounds.

It differs from standard quality assurance by focusing on adversarial critique rather than basic validation. It uses iterative review rounds, manifest-based bundling, and isolated iterations to converge on an Approved verdict for architecture decisions, security designs, or research findings through structured prompts.

You should avoid iterative red-team reviews for simple design approvals that do not require multiple review rounds or adversarial critique. This approach is built for complex architecture decisions, security designs, and research findings needing manifest-based bundling and isolated iterations to reach an Approved verdict.