Confirmation-Before-Action Prompt Pattern
Write a pattern that makes an agent show exactly what it is about to do.
Added October 15, 2026. Written with AI assistance and checked for clarity and placeholders; not reviewed by a lawyer, accountant, clinician or other professional. How these prompts are made
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Write a confirmation pattern for an assistant that can take actions such as [ACTIONS]. Which actions are reversible and which are not: [REVERSIBILITY]. What the user wants to see before approval: [APPROVAL_DETAILS]. Define a standard confirmation card with fields: action, target, exact content or amount, who is affected, whether it can be undone, and the two options Approve and Cancel (and Edit where relevant). Write the instruction that tells the agent when confirmation is required (anything that sends, spends, deletes, shares, or changes access), when it is not (reads, drafts), and what it must do if the user gives vague consent such as 'sure' to a bundle of actions (confirm one by one). Add rules for expiry of an approval and for repeated actions in a batch. Show confirmation cards for a payment, an email to an external recipient, and a file deletion, and a failing example that is too vague with a corrected version. Include a short list of user-interface tips to make confirmations readable on a phone.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.