fixme-handle-plan-review

You can classify plan-review findings by applying a fixed taxonomy—such as FIX, REJECT_FALSE_POSITIVE, or ASK_USER—while reading the actual repository code and plan context to verify each finding. This ensures findings map to consistent outcomes.

Taxonomy-driven finding classification is the process of routing plan-review findings into structured categories like FIX_UNCLEAR, REJECT_WONT_FIX, or REJECT_ALREADY_FIXED. It compares findings against locked decisions to enforce consistency and traceability.

Codebase verification reads the actual repository files and plan context to validate review findings against reality. It cross-references findings with locked decisions and decision logs to suggest an appropriate classification and generate routing instructions.

Yes, decision logging integrates with locked decisions and decision logs to enforce consistency and traceability. It ensures that classified findings align with previously established plan decisions and outputs structured data for downstream routing.

The best way to route rejected false positives is to use a structured taxonomy that includes REJECT_FALSE_POSITIVE and REJECT_WONT_FIX classifications. The system verifies findings against the codebase, logs the decision, and outputs structured routing instructions for downstream planning.

You should avoid automated finding classification when there are no locked decisions or decision logs available for cross-referencing, as the system relies on comparing findings against established plan context and repository code to generate accurate classifications.