Q14 of 60 — CCA Practice Exam

Claude Code for Continuous Integration

Your CI pipeline uses a programmatic tool-calling workflow where Claude invokes a run_test_suite tool to execute integration tests against a staging environment. The test suite takes 3–8 minutes to complete depending on the number of changed modules. Logs show that roughly 30% of CI runs terminate with a TimeoutError: Calling tool ['run_test_suite'] timed out in stderr, forcing Claude to retry and extending total pipeline time by 10–15 minutes per incident. The run_test_suite tool currently accepts a single module_list parameter and always executes all listed modules in one call. What is the most effective architectural fix?

A. Add a system prompt instruction telling Claude to retry immediately after a TimeoutError and limit retries to three attempts before escalating to a human reviewer, ensuring failures don't silently stall the pipeline.

B. Redesign run_test_suite to accept a bounded subset of modules per call and implement an internal timeout so each invocation exits cleanly within the container's execution window, letting Claude iterate across batches.

C. Increase the max_tokens budget so Claude has sufficient room to process the full test results before the session expires, preventing the container from being cut off mid-execution during long test runs.

D. Switch from programmatic tool calling to direct tool use so that each test module triggers a separate API round-trip, giving each call its own independent timeout window and isolating failures to individual modules.