review-researcher

Code review findings are independently re-investigated to verify their validity, trace root causes, and issue verdicts like confirmed, false-positive, partial, or needs-human with supporting evidence.

Root cause analysis for code review findings is automated by independently re-investigating each issue, tracing the underlying cause, and proposing minimal fixes with explanations for human decision.

Yes, fix suggestions are generated per issue after verification, providing minimal root-cause fixes with explanations so a human can decide whether to proceed with the correction.

Automated code review results undergo a truth check by dispatching a review-research phase that independently verifies each finding and outputs research notes with verdicts and recommended fixes.

Code review verification handles partial findings by issuing a partial verdict, identifying the true aspects of the issue, tracing the root cause, and proposing the appropriate fix.

Automated code review verification should not be used when findings do not require truth verification before a fix, or when the review-research phase is not explicitly dispatched for the task.