Replay compares a declared set of pinned decision inputs and expected receipt lineage against a new rule or receipt set.
Working method
- Freeze the corpus inventory before comparison and classify every item as agree, mismatch, or not-comparable.
- Review mismatch inputs as behavior changes; do not regenerate old receipts to erase history.
- Preserve runner identity, rule bytes, input bytes, and result observations for reproducibility.
Boundaries
- Not-comparable is a disclosed boundary, not a mismatch silently counted as agreement.