A housing authority's case-management system already has a prebuilt MCP Server that covers case creation, status updates, and document attachment. The architect chooses to connect the agent through that prebuilt MCP Server rather than building a custom API integration against the case-management system directly. When is this the sounder choice?
Select an answer to reveal the explanation.
Short Explanation
Think of a prebuilt MCP Server like a paved road that already goes exactly where you need to go: if it already covers case creation, status updates, and document attachment, building a parallel road alongside it just means twice the maintenance for the same destination. Reaching for the prebuilt server when it already covers the operations the agent needs is the sounder call - the other options describe situations that don't actually favor a custom build either.
Full Explanation
The case for using a prebuilt MCP Server over a custom API integration rests on coverage: when the prebuilt server already exposes the operations an agent needs - here, case creation, status updates, and document attachment - building an equivalent custom integration duplicates work that's already been done and needs to be independently maintained going forward. That's a straightforward efficiency argument, not a fallback of last resort. A case-management system with no standard API and only undocumented internal calls actually argues against relying on a prebuilt MCP Server, since a prebuilt server typically wraps a stable, documented interface - undocumented internal calls are a fragile foundation regardless of which integration approach wraps them. Whether the case data shares a common data model with other external systems doesn't bear on this choice at all; data-model overlap is a separate question about downstream schema mapping, not about whether an existing MCP Server already covers the operations needed here. And needing only read access to case status doesn't change the underlying logic - if a prebuilt MCP Server covers that read operation, the same efficiency argument for using it still applies; the operation being narrower doesn't tip the decision toward building something custom. The scope caveat: this reasoning holds only as long as the prebuilt server's operation coverage genuinely matches what the agent needs; a gap between what it exposes and what the workflow requires would reopen the custom-integration question. A concrete check: confirm the prebuilt MCP Server's documented operations cover every case-management action the agent's workflow actually calls before ruling out a custom build.