progressive-harsh-review

An adversarial multi-persona review evaluates non-code deliverables like product specifications by deploying three critic personas to score correctness, simplicity, testability, edge cases, and security, aggregating weighted means to enforce hard vetoes for critical defects.

A hard-veto rule in design review scoring automatically fails an artifact if critical defects are detected across weighted dimensions like security or correctness, ensuring stakeholders never see flawed plans.

Yes, you can apply a multi-persona critic model to review strategic plans before presentation, utilizing specialized personas to independently flag edge cases, echo-reasoning, and shared blind spots in the document.

Correlation detection during a harsh design review flags echo-reasoning and shared blind spots by cross-referencing the individual evaluations of multiple critic personas, highlighting areas needing deeper analysis.

Multi-persona scoring works effectively for non-code written artifacts and design documents by applying standardized evaluations across correctness, simplicity, testability, edge cases, and security before stakeholder presentation.

You should avoid adversarial weighted aggregation review for simple, low-stakes artifacts that lack complexity, as the rigorous multi-persona process is designed to catch hidden flaws in detailed specifications, designs, and strategic plans.