Question
Greenfield Go service with an AI agent: a scheduled-task runner that executes recurring jobs (cron expressions) like 'send the daily digest at 9am' and 'reconcile invoices hourly', deployed across 3 replicas for HA. Describe the build plan: how you parse/evaluate schedules, ensure each run fires once despite multiple replicas, handle missed runs, and the acceptance criteria. What does a careless 'run these jobs on a schedule' prompt get wrong in a replicated deployment?
Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it — tests, edge cases, reading critically — and how you’d re-prompt or decompose to get it right.
Vibe coding: describe the solution in plain language (or narrate it) and the coach grades your approach. Generating runnable code from your description is coming next.