FREE MARKDOWN TEMPLATE · NO ACCOUNT

AI agent handoff template.

Turn a finished agent run into a handback a human can actually decide from: what was requested, what the record supports, what remains a claim, and what needs attention next.

Download Markdown Preview the template

A static file. No signup, form, or imported activity.

A USEFUL HANDOFF KEEPS THESE APART

OBSERVEDtest command exited 1

REPORTED“all tests pass”

UNKNOWNno later test run is present

NEXT HUMAN DECISIONRerun checks before approving a push.

WHY A HANDOFF?

A completion message is not a review record.

An agent summary often collapses actions, claims, failures, and missing checks into one confident paragraph. A handoff preserves the distinctions a responsible human needs before allowing the next consequential action.

COPY, ADAPT, OR DOWNLOAD

The evidence-labelled handoff.

Complete this from the underlying activity record, not from the final summary alone.

# AI agent handoff

## Run
- Task:
- Date and time:
- Reviewer:
- Source record:

## 1. Requested
What outcome did the human ask for?

## 2. Authorised
What was allowed, reserved, denied, or unclear?

## 3. Executed and changed
| Label | Record reference | What happened or changed |
| --- | --- | --- |
| Observed |  |  |
| Governed |  |  |
| Prevented |  |  |

## 4. Claims and gaps
| Label | Claim or gap | Why this label applies |
| --- | --- | --- |
| Reported |  |  |
| Inferred |  |  |
| Unknown |  |  |

## 5. Verified
- Independent check:
- Actual result:
- Missing evidence:

## Next human decision
- Decision: review / approve / approve with conditions /
  retry / investigate / stop
- Reason:
- Conditions:

Download the free template

WORKED EXAMPLE

The agent says “done.” The handoff says what that means.

OBSERVED

src/auth/session.ts was modified. The recorded test command exited 1. A later edit targeted the test file.

REPORTED

The agent's final message says authentication is fixed and all tests pass.

UNKNOWN

No later successful test execution is present in the supplied record.

NEXT DECISION

Rerun the relevant checks and inspect the diff before approving a push.

TRUST BOUNDARY

Useful evidence is still bounded evidence.

A supplied log does not prove omitted events did not occur.
A receipt identifier remains a reference until checked against its canonical store.
A successful tool call does not by itself prove the intended outcome.
This template is a review method, not legal, regulatory, certification, or compliance advice.

HAVE A LOG NOW?

Generate the first draft in your browser.

Traceplain turns JSON or NDJSON activity into an evidence-labelled handback. Imported activity stays on your device.

Open the free reviewer

NEED THE FULL METHOD?

Get the Founding Review Kit.

Eight printable pages, approval gate, seven-day team trial, worked example, Markdown template, and JSON Schema.

Get the kit · A$29 once