Dbs Good Question

This turns fuzzy questions into structured problem briefs that AI agents can actually reason about. Instead of "why isn't my content working?" you get a breakdown with observable phenomena, core conflicts, constraints, and testable hypotheses. It scores your question across five dimensions (object, goal, conflict, constraints, feedback) and flags what's missing before generating candidate explanations. The philosophy is that good questions compress search space and expose contradictions, bad ones force agents to rely on default assumptions. Works through triggers like /好问题 or asking "can an agent solve this" in Chinese or English. Most useful when you know something's wrong but can't articulate what you're actually trying to figure out.