CCA-F: Claude Certified Architect — Foundations

By Anthropic · Question Mode

QuestionQ127

Prompt Engineering & Structured Output

You are incorporating Claude Code into your Continuous Integration/Continuous Deployment (CI/CD) pipeline. The system performs automated code reviews, creates test cases, and gives feedback on pull requests. You must design prompts that deliver actionable feedback while minimizing false positives.

Your automated code review is failing to report genuine bugs in pull requests. An investigation shows that the review prompt says: “Only flag critical issues that would definitely cause production failures. Ignore minor concerns and anything you're uncertain about.” Developers confirm that some missed bugs are real logic errors that the model investigated but decided not to report. The team requires review output to remain structured, with each finding tagged with metadata, and actionable.

Which prompt modification both eliminates the cause of the suppressed findings and retains structured, tagged output for downstream filtering?

Explanation

A reporting policy that requires certainty and production-critical impact suppresses valid findings by design. Emitting all findings with confidence and severity metadata preserves each potential bug for review while allowing a downstream system to apply its own thresholds and filters. Claude Code supports structured JSON output for programmatic automation.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!