Write a one-off backfill script with a dry-run mode
A one-off data fix delivered as a reviewed script with a dry run, with execution left to a person.
The ticket
Write a backfill script with a dry-run mode and attach its output.
Acceptance criteria
- A dry-run mode reports what would change without changing it
- Dry-run output is attached to the pull request
- It processes in batches and can resume after interruption
- Execution stays with a person, not with this ticket
What lands as proof
Dry-run output showing the exact rows and values that would change, reviewed before anything runs.
Why teams defer it
- A backfill that goes wrong is expensive and hard to reverse, so it needs a reviewer who understands the data.
- Writing it to be resumable takes longer than writing it to run once, and the difference only matters when it fails halfway.
Questions
- What does the agent actually change?
- The ticket is scoped to one outcome: write a backfill script with a dry-run mode and attach its output. Work that serves that outcome is in scope, and anything outside it is left for a separate ticket, so the pull request stays reviewable.
- How do I know the work is done?
- The pull request carries the evidence, not only the diff. Here that means dry-run output attached, the execution stays with the team, so a reviewer can confirm the result without reproducing the work locally.
- How much oversight does this need?
- The run stops for the plan, and the review bar is raised afterwards so that smaller suggestions block the merge too, not only serious findings.