Your AI code review system supports team-specific guidelines, and one customer has about 800 rules covering style, security, and framework conventions across several languages. Putting all of them in every prompt is costly, and the model applies Python rules to Go files while ignoring most of the others. What should you do?

1. A Put all 800 guidelines into every review prompt, since a long context window lets the model pick the rules that apply to each diff.
2. B Drop the custom guidelines entirely and rely on the model's general knowledge of best practices, which covers most team conventions.
3. C Fine-tune a review model on the guideline document so it internalizes all 800 rules and needs no rules in the prompt at review time.
4. D Index the rules, retrieve the few relevant to each chunk's language, paths, and APIs, and cite the matching rule in each comment.