Q27 of 60 — CCA Practice Exam

Developer Productivity with Claude

A developer productivity team has deployed Claude to automate configuration updates across a large internal project — updating 120 YAML configuration fields based on a change manifest. After two weeks in production, the team reports that roughly 8% of update runs result in silent failures: fields that were supposed to be updated are unchanged, or worse, two related fields are set to conflicting values. Logs confirm Claude is completing its runs without errors. What architectural change would most effectively catch these issues before they corrupt the configuration files?

A. Add explicit instructions to the system prompt telling Claude to double-check every field reference against the manifest before applying any updates to the YAML files.

B. Implement a plan-validate-execute pattern: have Claude write intended changes to an intermediate changes.json file, then run a validation script against it before applying updates.

C. Increase the context window allocation so Claude has more token budget to reason over the full set of 120 fields simultaneously before committing any changes to the YAML files.

D. After each run, trigger a separate Claude pass to review the updated configuration files and flag any suspicious field values or conflicting entries for manual inspection.