LOGIC OR NOT P067 — FORMAT FINDING & REPAIR SCHEMA — v0.1.2


review_id:
artifact_ref:
artifact_version_or_as_of:
scope:
observed_at:


baseline:
  baseline_type: EXPLICIT_STYLE_GUIDE | USER_INSTRUCTION | ESTABLISHED_DOCUMENT_PATTERN | INFERRED_PREFERENCE | NONE
  baseline_ref:
  authority_level:
  baseline_limit:


observability:
- surface: STRUCTURE | TEXT | STYLES | LAYOUT | TABLES | NUMBERING | RENDER | OTHER
  status: OBSERVED | PARTIAL | NOT_OBSERVABLE
  evidence_or_method:


findings:
- finding_id:
  finding_type: BASELINE_INCONSISTENCY | INTERNAL_INCONSISTENCY | LAYOUT_OR_RENDER_DEFECT | ACCESSIBILITY_OR_USABILITY_ISSUE | VARIATION_ALLOWED | NOT_OBSERVABLE | SEMANTIC_RISK
  severity: BLOCKING | MATERIAL | MINOR | NOTE
  statement:
  artifact_locator:
  baseline_ref:
  evidence:
  semantic_risk:
  proposed_repair:
  repair_status: PROPOSED | APPLIED | VERIFIED | NOT_APPLIED | HOLD_SEMANTIC_RISK
  repair_artifact_ref:
  verification_ref:


verification:
  semantic_content_unchanged: PASS | FAIL | UNKNOWN
  baseline_conformance_after_repair: PASS | FAIL | NOT_APPLICABLE | UNKNOWN
  render_or_layout_check: PASS | FAIL | NOT_OBSERVABLE | NOT_APPLICABLE
  unresolved_findings: []
  limits: []


RULES
No objective inconsistency claim from INFERRED_PREFERENCE alone. APPLIED requires real edit evidence; VERIFIED requires post-edit readback. SEMANTIC_RISK is never auto-repaired as format-only.