Your code review bot must enforce 300 team-specific style and security rules. Pasting all 300 into every prompt is slow and expensive, and the model ignores many of them. What is the better design?

1. A Fine-tune the model on all 300 rules once, then never include any rule text in review prompts.
2. B Retrieve only the rules relevant to each diff chunk, by language and path.
3. C Keep all 300 rules in every prompt but bold the key ones so the model weighs them more heavily.
4. D Drop the custom rules and rely on the model's general knowledge of best practices per language.