Designing content for high-stakes moments - Pro
A working reference for the delete, the cancel, and the actions a user can't take back. Eighteen pages, a quick-reference card, and a Claude skill that evaluates your confirmations against the same system.
Most teams handle destructive actions one at a time. A modal gets added because someone asked for a warning. A stakes sentence gets written from scratch, or copied from the last one and left slightly wrong. The pattern only becomes visible after something ships that shouldn't have.
This reference treats those moments as a system. It separates the container, which governs whether the user stops, from the message, which tells them why. A four-question decision tree returns the friction level an action warrants, and five worked states show what that looks like in real UI, with the before-and-after for each. What's inside
The 18-page guide and the quick-reference card
The Pro Claude skill: evaluates a whole flow at once instead of one dialog at a time.
A flow-level anti-pattern check that only works across a set: it catches consent fatigue, inconsistent friction for equivalent actions, and the problems a single-dialog review cannot see.
Direct write into your Notion working file, so the skill's audit lands in your spec.
Core or Pro. Core walks one confirmation at a time and hands you the friction level and the fix. Pro reads a whole flow at once, catches where your dialogs disagree with each other, and writes the audit into your Notion. Core for the occasional dialog; Pro for designing destructive actions across a product.
