from-the-other-side-anitta

Use structured challenge prompts that ask what is being assumed about data quality, causality, and stability. Then test counterfactuals by asking what observation would change the conclusion, and check whether each reasoning step follows logically.

Match your language to evidence strength: use 'suggests' for weak correlations, 'indicates' for moderate support, and 'demonstrates' only for strong, replicated evidence. Avoid claims stronger than the available data supports.

Use it when the cost of being wrong is meaningful, such as high-stakes decisions, leadership presentations, or analyses with significant uncertainty. It is less necessary for low-stakes exploratory work where speed matters more.

Challenging reasoning tests whether logic steps follow and assumptions hold. Challenging interpretation examines whether the narrative and framing accurately represent the evidence. This profile addresses reasoning first, interpretation second, and rigor checks third.

Yes, it is designed to coordinate with complementary profiles: one drives momentum and deliverables, another refines narrative clarity, while this one validates that underlying reasoning holds. Handoffs occur when uncertainty in assumptions arises.