Q39 of 60 — CCA Practice Exam

Developer Productivity with Claude

Your team has integrated a custom SearchCodebase tool into a Claude-powered developer productivity agent. The tool performs read-only semantic search across your monorepo and never modifies any files. After deploying, developers report that multi-file investigation tasks — where Claude fans out to search several modules simultaneously — are taking 40–60 seconds per request. Examining the agent logs, you see Claude is issuing 6–8 SearchCodebase calls per task, but each call starts only after the previous one completes. What is the most effective fix?

A. Increase the max_tokens budget for the agent so Claude has more room to issue all search calls in a single response turn instead of spreading them across multiple turns.

B. Mark SearchCodebase as read-only in its tool annotations (readOnly in TypeScript or readOnlyHint in Python) so the SDK can execute multiple instances concurrently rather than sequentially.

C. Add SearchCodebase to allowed_tools / allowedTools so the SDK auto-approves each call without waiting for a permission prompt before starting the next one.

D. Rewrite SearchCodebase as a single batched tool that accepts an array of search queries so Claude can express all lookups in one tool call.