Q55 of 60 — CCA Practice Exam

Claude Code for Continuous Integration

Your CI pipeline uses claude_agent_sdk to automatically diagnose and fix failing tests in a monorepo. The agent is configured with allowed_tools=["Read", "Edit", "Bash", "Glob", "Grep"] and max_turns=30. After three weeks in production, engineers report that roughly 15% of CI runs are silently completing with no code changes made, even when the test failures are reproducible and fixable. Examining the logs, you find these runs consistently produce ResultMessage subtypes of error_max_turns. The failing tests are in deeply nested service modules that require extensive file traversal before any fix can be applied. What is the most effective architectural response?

A. Add "Bash(find *)" and "Bash(tree *)" as additional scoped tool permissions to speed up file traversal and reduce the number of turns consumed before the agent reaches the relevant files.

B. Handle error_max_turns in the result loop by resuming the session with an increased max_turns limit, and add a disallowed_tools list excluding broad commands like Bash to force the agent to use more targeted read-only traversal tools.

C. On error_max_turns, capture the session_id from the ResultMessage and resume the session rather than silently completing, so the agent continues from where it stopped instead of discarding progress.

D. Switch the agent to effort="low" so it uses fewer turns on reasoning overhead, freeing up the remaining turn budget for actual file edits and test fixes.