Q50 of 60 — CCA Practice Exam

Multi-Agent Research System

Your multi-agent research system includes a content moderation layer that screens research summaries before they are passed to a synthesis agent. Analysts report that roughly 8% of legitimate research findings about financial markets are being flagged and dropped — particularly comments like "commodities are outperforming equities this quarter" — while some genuinely problematic speculative investment claims pass through undetected. The current moderation prompt lists Specialized Advice as a prohibited category with no further definition. What change would most effectively reduce both the false-positive and false-negative rates without adding a separate review agent?

A. Replace the flat prohibited-category list with a category-to-definition dictionary, pairing each category name with a precise description of what content it covers — for example, specifying exactly which financial claims constitute prohibited Specialized Advice.

B. Increase max_tokens for the moderation call so Claude has more room to reason through ambiguous financial content before producing its classification decision, reducing the chance of a hasty or truncated judgment.

C. Add a second moderation pass that re-evaluates every flagged item, instructing the second pass to "be more conservative and avoid false positives" to catch errors the first pass missed.

D. Expand the prohibited-category list by adding finer-grained subcategories such as Investment_Advice and Market_Commentary, relying on Claude to infer the boundary between them from the subcategory names alone.