CCAR-F scenarios › Multi-Agent Research System · Question 24 of 60
Scenario 3 · Multi-Agent Research System
You're building a research system on the Claude Agent SDK in which a coordinator agent delegates work to specialized subagents: one searches the web, one analyzes documents, one synthesizes findings, and one generates reports. Given a topic, the system produces a thorough report with citations.
Agentic Architecture & Orchestration
On the topic "remote work and urban real estate," the coordinator sends the same broad brief to three web search subagents. Their combined results contain the same dozen widely cited articles three times over, while commercial office vacancy data, residential migration, and municipal tax effects are barely covered. What change best improves coverage?
Show answer and explanation
Answer: D — Partition the research scope across the subagents, assigning each a distinct subtopic or source type, such as office vacancy data, residential migration, and municipal finance.
Giving every subagent the same brief produces duplication; partitioning scope by distinct subtopics or source types minimizes overlap and spreads coverage (task 1.2). Deduplication in synthesis (a) tidies citations but can't add coverage that was never gathered. More subagents with the same brief (b) multiply the duplication. More results per query (c) adds cost and noise without directing any effort at the gaps.
More from this scenario
- To keep the configuration simple, every subagent's AgentDefinition grants the same 16…
- The document analysis subagent has one tool, process_document(url, mode), described as…
- One of eight sources the web search subagent is retrieving returns an HTTP 503 once. The…
- Two credible sources give different figures for the same metric: an industry association…
- The coordinator concatenates outputs from six subagents, about 60 pages, and passes them…