# Automation handover and completion record

Fill in the handover once, then use the run record for significant executions and exceptions. Keep credentials in approved storage; reference their location without copying them.

## Handover

**Workflow / owner / alternate:** ____________________

**Business purpose and who depends on the output:** ____________________

**Trigger (schedule, event, manual) and expected frequency:** ____________________

**Input scope / intended outcome:** ____________________

**Systems read / systems written, with the identity used for each:** ____________________

**Unrelated state that must remain unchanged:** ____________________

**Version / configuration / runbook references:** ____________________

**Where failures are reported and who is notified:** ____________________

**Stop and recovery procedure:** ____________________

**Credential storage location, owner, and rotation date:** ____________________

**Change procedure (who approves a change, where it is tested, how it is rolled back):** ____________________

## Run record

| Run ID / time | Attempted action | Destination response | Authoritative verification | Exception | Next owner / action |
| --- | --- | --- | --- | --- | --- |
|  |  |  |  |  |  |

## Retry rule

**Duplicate consequence:** ____________________

**Supported duplicate protection or pre-check:** ____________________

**When to stop and escalate:** ____________________

## Second-person exercise

- [ ] Locate the current version and configuration.
- [ ] Perform an ordinary run.
- [ ] Explain its completion evidence.
- [ ] Recognize and resolve a known failure.
- [ ] Resume without duplicating completed work.
- [ ] Stop the workflow and identify affected partial work.
- [ ] Find the credential location and confirm who can rotate it.
- [ ] State what would happen if the workflow silently stopped running for a week.

**Observed handover gaps and owners:** ____________________

**Log access / retention:** ____________________

## Health check

| Question | Answer | Evidence |
| --- | --- | --- |
| Did every expected run in the period happen? |  |  |
| Were all failures noticed by a person within the agreed time? |  |  |
| How many manual corrections were needed? |  |  |
| Are any credentials or interfaces due to expire before the next review? |  |  |

## Common failure modes

- The automation runs under the builder's personal account and stops when they leave.
- Failures are logged but nothing tells a person.
- A retry after a timeout creates a duplicate the destination cannot detect.
- The runbook describes the original version, not the one in production.

**Review cadence / next review:** ____________________

**Retirement trigger and who decides:** ____________________

This record covers operability and handover. It does not judge whether the workflow should exist; use the complexity review for that.