super-review:audit-self

To reduce false positives in GitHub PR reviews, you can audit historical review findings against merge outcomes to identify dismissed or quietly ignored patterns. This allows you to propose configuration overrides and allow-lists that align automated checks with your team's actual codebase conventions.

Review calibration drift occurs when automated PR analytics diverge from actual codebase conventions over time. Auditing prior findings against recent merge outcomes detects this drift, enabling targeted prompt edits and severity adjustments to realign automated governance with repository reality.

You can audit AI code review history by aggregating past PR comments and classifying them by merge outcomes like confirmed, dismissed, or outdated. Mining these patterns by sub-skill and anti-pattern slug reveals actionable proposals for SKILL.md prompt edits and new red flag detection rules.

Yes, PR analytics can detect missed positive bugs by evaluating review findings against actual repo history. When recurring real bugs are not yet covered by existing checks, the audit recommends new redFlagAddons and targeted prompt edits to expand detection coverage.

The best way to configure severity overrides is by mining high false-positive and high-confirm patterns from recent PR history. Analyzing these patterns by anti-pattern slug allows you to generate concrete configuration changes for your JSON settings to reduce noise and focus on real issues.

Automated code review keeps flagging the same false positive patterns because its rules lack alignment with your specific team conventions. Auditing prior findings against merge outcomes identifies these recurring patterns, allowing you to apply allow-lists and configuration overrides to stop the noise.