refactor-classification-policy

RTL refactoring changes are classified into SAFE, RESTRICTED, or PROHIBITED tiers to determine approval and gating requirements. This categorization eliminates uncertainty by enforcing verification rules and escalation policies that prevent behavioral regressions or contract violations.

SAFE tiers cover behavior-preserving cleanups, RESTRICTED tiers involve structural or test-affecting edits requiring approval, and PROHIBITED tiers involve semantic, timing, or interface contract changes. Each tier enforces specific pass/fail gates and escalation policies for automated review workflows.

Approval gates for ambiguous RTL edits are enforced by marking risky or unclear changes for manual review during continuous integration. The classification policy requires explicit approvals for prohibited changes and demands equivalence confidence for SAFE and RESTRICTED modifications.

Yes, you can generate audit trails for RTL refactoring workflows by producing a structured Refactor Execution Report. This report captures classification results, approval statuses, and gating decisions for downstream automation, CI gating, and audit purposes within RTL design pipelines.

Automating RTL code review gating for commit-level change sets is best handled by applying safety tier classification directly to the proposed edits. This process enforces pass/fail gates based on the SAFE, RESTRICTED, or PROHIBITED designation of each change set.

Equivalence confidence should be required for SAFE and RESTRICTED refactoring changes during automated review workflows. PROHIBITED changes bypass this by enforcing strict approval rules, while ambiguous or risky edits are escalated for manual review.