agency-minimal-change-engineer

To prevent scope creep during code generation, apply a minimal change engineering approach that enforces line-by-line justification, ensuring the AI only outputs the smallest diff required to satisfy the specific task requirements.

The best way to keep code diffs minimal for high-quality PR reviews is to enforce surgical coding discipline, rejecting unrequested refactoring and defensive coding patterns so that every line added is strictly required by the task.

To stop AI from making large refactoring changes when you only need a bug fix, apply a strict scope boundary protection persona that rejects over-engineering and guides the AI to provide a one-line patch rather than a 400-line cleanup.

Yes, you can use minimal change engineering to reduce technical debt in your software development workflow by enforcing a disciplined approach that avoids "while I'm here" refactoring, which often leads to regression bugs and unnecessary scope expansion.

You should not use a surgical coding approach when a task explicitly requires broad refactoring, architectural changes, or defensive coding patterns, as this discipline strictly limits code modifications to the smallest possible diff to satisfy immediate requirements.

No, minimal change engineering does not require any specific dependencies or components; it works by applying a disciplined engineering persona to your current coding session to ensure all future diffs are strictly limited to task requirements.