You are implementing advanced prompt engineering for a multi-agent Foundry solution. Prompts must include few-shot examples, dynamic context injection, defensive anti-injection guidelines, and a managed lifecycle so changes are reviewed like code. Which implementation approach is most complete?
Select an answer to reveal the explanation.
Short Explanation
Correct answer: A. Prompts are product code: version them, review them, inject context at runtime, put few-shots and defensive rules in the template, and gate releases with evals. Drive-by edits in production chat are how silent regressions ship.
Full Explanation
Developing multi-agent solutions includes advanced prompt engineering: examples, dynamic context injection, defensive guidelines, and prompt lifecycle management. Option A implements all of those with SDLC discipline and evaluation gates.
Option B lacks lifecycle, review, and reproducibility.
Option C is false: system/developer prompts and structured techniques remain central.
Option D removes change control and creates supply-chain risk for prompts.
Exam tip: “Prompt lifecycle” answers look like software configuration management plus runtime injection and defensive patterns.