Review a Set of AI-Generated Code Changes
Reviews a diff by real-world risk instead of file order and prioritizes defects that can cause failures.
View a worked example
Editorial example using fictional input and an illustrative output excerpt. It is not a recorded model test.
Example input
JavaScript diff: - return items.length ? total/items.length : 0; + return total/items.length; Existing tests only use items=[1,2].
Example output
The change removes the empty-list guard. Location: return line. Trigger: items=[], total=0. Result: NaN. Tests cover only nonempty inputs. Minimal fix: restore the empty branch. Add tests for empty lists returning 0 and unchanged nonempty averages.
Check the result
- Run the empty-list reproduction.
- Tie the defect to the changed line, not a style preference.
- Confirm that the business contract requires 0 for an empty list.
Fill in the blanks0/1
Fill these in and the prompt updates as you type. Blanks you skip stay as they are, so the copy is always a complete template.
Prompt
You are a senior engineer responsible for release quality. Review the provided [code changes or diff] and look for defects with concrete triggers that affect correctness, security, performance, or compatibility. First explain what the change is trying to accomplish. Then list findings by severity. Every finding must include the affected location, trigger, user-visible impact, why existing tests missed it, and the smallest reasonable fix. Do not report style preferences as defects or repeat the same root cause. Finish with the tests that matter most, the parts that are safe to merge, and the questions that must be answered first. If you find no clear defect, say so and list the remaining uncertainties.
Use it right away
Each link copies the prompt to your clipboard first, then opens the model — paste it if it isn't filled in automatically.