# Findings for round 001

Evidence status: illustrative. These findings describe the authored walkthrough.

| Finding | Sample | Failure mode or issue | Disposition |
|---|---|---|---|
| F01 | D1 | Axis label promoted to title | Amend R02 to use visual role and preserve the document label separately. |
| F02 | D2 | Standard error treated as a variable | Amend R03 to attach the statistical form where the layout supports its association. |
| F03 | D2 | Meaning of stars is unclear from the data snapshot | H01 keeps the visible note and leaves thresholds unspecified. Gold and schema stay at the pinned versions. |
| F04 | C2 | Counted entity inferred from an unlabeled scale | Add R05. The intended entity remains unspecified and the rule needs testing on fresh samples. |

| Rule-change tally | Count | Rule IDs |
|---|---:|---|
| Added | 1 | R05 |
| Amended | 2 | R02, R03 |
| Retired | 0 | None |
| Active in P002 | 5 | R01–R05 |

The accepted rule-change count is `C_001 = 1 + 2 + 0 = 3`. C1 had no flagged finding. H02 records a selected human spot check. The council discarded generic title substitution and the blanket ban on units when an axis lacks a unit label. The orchestrator carried F04 forward as an open validation question. This single illustrative transition shows the accounting method. A stopping decision would require observations from additional batches and human judgment.
