One PR-review category floods the clerk portal team with false positives while its prompt is being rewritten. What should they do meantime?
Select an answer to reveal the explanation.
Short Explanation
If one smoke detector chirps every five minutes, you take out the battery while you fix it—not the whole fire system. Park the noisy category until the prompt is rewritten.
Full Explanation
When one PR-review category floods the clerk portal team with false positives while its prompt is being rewritten, the correct interim control is to temporarily disable that noisy category and keep the rest of review running. Parking the broken detector restores trust and developer attention while the prompt is fixed, without abandoning security or other healthy categories.
Leaving the noisy category on so trust can keep falling fails because continued noise accelerates ignore behavior and may cause teams to disable the whole bot later under pressure. Disabling all review categories including security for the entire organization fails as over-correction: one bad category does not justify removing real vulnerability detection citywide. Increasing the volume of noisy comments to force developers to adapt fails psychologically and operationally—adaptation to spam is numbness, not better security.
Exam caveat: temporary disable is a mitigation while rewriting; it is not permission to leave the category off indefinitely without a fix or replacement control. Operational check: feature-flag or prompt-gate the noisy category off in CI, announce the window, ship the rewritten prompt with examples, then re-enable on a canary set of clerk-portal PRs and measure false-positive rate before full restore.