# AI handoff readiness checklist

Use this checklist before another human or AI agent relies on a handoff.

## Required fields

- [ ] **Objective:** the intended outcome is explicit.
- [ ] **Current state:** completed, active, blocked, and unverified work are distinguishable.
- [ ] **Evidence:** material claims point to a file, URL, command, query, result, artifact, or exact source location.
- [ ] **Next action:** one bounded action starts with a concrete verb and names an observable result.
- [ ] **Ownership:** a person, agent profile, service identity, or team owns the next action.

## Supporting fields

- [ ] **Acceptance:** success can be observed or measured.
- [ ] **Decisions:** material choices include rationale, or the record explicitly states that none exist.
- [ ] **Constraints:** permissions, safety boundaries, exclusions, deadlines, budgets, and dependencies are visible.
- [ ] **Receiver verification:** the receiver must inspect current evidence or reproduce one material result before acting.

## Verification boundary

Passing this checklist does not prove that a claim is true, evidence is current,
access is sufficient, or the next actor can safely perform the work. The
receiving identity must inspect the cited sources and current revision before
continuing.

Interactive browser checker:
https://handover.sh/tools/ai-handoff-readiness-checker

Open Handoff Continuity Record:
https://handover.sh/protocol