Prompt Engineering & Structured Output (D4)
Prompting techniques, structured outputs, JSON schema, few-shot.
Prompt Engineering & Structured Output (D4) is the craft of getting precise, machine-usable answers out of Claude: explicit criteria, few-shot examples, and JSON-schema-enforced output you can trust downstream. The exam tests whether your prompts hold up under real data, not just happy-path demos.
Expect questions on designing prompts with explicit criteria to cut false positives, few-shot prompting for consistency, enforcing structured output with tool use and JSON schemas, building validation, retry, and feedback loops for extraction quality, efficient batch processing, and multi-instance or multi-pass review architectures.
This domain spans 6 subtopic areas, covered by 126 practice questions across 31 easy, 65 medium, and 30 hard items.
Your CI/CD pipeline posts Claude Code feedback as PR comments, but reviewers complain that 30% flag valid code as problematic. What prompt change most directly reduces these false positives?
What's covered
Subtopic areas in Prompt Engineering & Structured Output, drawn from the exam blueprint: