A handoff card for collaborating with an AI twin

An AI twin could make collaboration easier by carrying a clear handoff card into each task. Here is a proposed format to test—not a feature I am claiming is already built.

1. Goal: What should the other person be able to decide or do after this contribution?
2. Evidence: Separate facts supplied by the person, assumptions, and open questions.
3. Permission: State what the twin may draft and which actions need the person's approval.
4. Credit: Preserve who supplied the original idea and who changed it.
5. Exit: Let either collaborator pause, decline, or request a human without losing their work.

Example: two people want to turn a rough idea into a learning session. The twin drafts a 20-minute outline from their agreed notes, marks two unresolved choices, and returns it for review. It does not send invitations or describe either person as committed.

A small test: give two pairs the same fictional task, one with this card and one with an ordinary brief. Compare clarification questions, incorrect assumptions, and whether each person can explain what was approved. This would be exploratory, not proof of effectiveness.

The trade-off is extra setup. A one-sentence card may suit a casual task; shared or consequential work needs more detail. Which field would you simplify first, and what would you keep even in the shortest version?