design:review

An adversarial design review evaluates technical proposals by simulating expert agents to critique architecture, security, and operational concerns. It iteratively refines the design through proposal, critique, and revision loops while incorporating human reviewer feedback.

An adversarial critique loop is a process where simulated agents representing different expert perspectives evaluate and challenge an architectural design. The loop facilitates iterative refinement of the technical proposal by cycling through design generation, critique, and human-directed revisions.

Yes, you can refine an existing architectural design document. The review process supports starting from either a new topic or an existing document, applying simulated expert critiques and iterative refinement to enhance the original technical proposal.

You export a finalized technical proposal by selecting either DOCX or Markdown formats. After completing the adversarial critique loop and incorporating human feedback, the skill generates the refined architectural design document in your chosen format.

Yes, the design review process supports human feedback during iterations. It orchestrates a loop of proposal and critique that allows human reviewers to provide oversight, direct the next iteration, and ensure comprehensive evaluation of feasibility and security concerns.

Simulated agents provide automated multi-perspective critiques but require human oversight to ensure comprehensive evaluation. The simulated architect and quality advocate identify risks, but human reviewers must direct iterations to validate feasibility and address complex operational concerns.