Your company already runs a mature inventory agent. A new multi-agent fulfillment solution must use it without rewriting it, while also attaching new MCP tools for carrier APIs. Which integration approach is best?
Select an answer to reveal the explanation.
Short Explanation
D is the exam-aligned reuse path: bring existing agents in through A2A and/or MCP, and extend capabilities with new MCP tools—securely, not by pasting keys into prompts. A leaks secrets. B wastes a working asset. C is manual swivel-chair integration. Multi-agent development explicitly includes integrating existing agents with A2A/MCP.
Full Explanation
D is correct because developers should integrate existing agents into multi-agent solutions securely using A2A and/or MCP, while building tool ecosystems for new capabilities. A is incorrect: distributing API keys via prompts is insecure. B is incorrect: unnecessary rewrites delay delivery when a stable agent exists. C is incorrect: manual UI copy is not orchestration. Define auth, contracts, and error handling at the integration boundary; treat the legacy agent as a bounded capability.