A team updates system prompts for three specialist agents weekly. Last month a prompt edit improved tone but increased incorrect tool arguments. You need an evaluation practice that prevents similar regressions. What should you do?
Select an answer to reveal the explanation.
Short Explanation
C is solid release hygiene for prompts. Prompt lifecycle management in multi-agent systems means versioning plus automated evals on a frozen scenario pack covering tool accuracy, task success, and safety—before promotion. Author preference and doc-only review miss runtime regressions like bad tool arguments. Unversioned prompts in chat history are operational chaos. Treat prompts as deployable artifacts with regression gates, same spirit as code.
Full Explanation
Correct Answer — C
Prompt lifecycle management and evaluations for prompts require versioned prompts, automated offline evals on representative scenarios (including tool-call correctness), and promotion gates. That catches tone-improving edits that break tool arguments.
Why A is wrong: Preference-based production pushes lack regression control.
Why B is wrong: Static reading does not exercise tool calls or multi-agent runtime behavior.
Why D is wrong: Unversioned prompts in chat history prevent controlled rollout and rollback.
Exam tip: Version prompts + automated scenario pack evals before promote.