sdlc-intentlisted

# Capture the intent First file in the chain (`ai-native-sdlc`). It exists so that an idea stops waiting on someone with the right job title to write it up, and so what reaches implementation still resembles what the originator actually meant. ## Is it worth writing? **Yes** when a second person would need it written down to act, or when someone six months out will ask why this was built. **No** for a typo, a rename, a one-line fix, or anything where the write-up costs more than the work. Judgment call, and erring toward "no" is cheaper than erring toward ceremony. The originator can be anyone — the person with the idea, whoever triaged the ticket, or an agent that noticed a metric move (`sdlc-close-loop`). The route in doesn't change what gets written. ## Interview before template The failure mode is reaching for the template immediately and filling it with the vague version of the idea. Let them describe the problem however they describe it. What can't they do today? Who else hits this? What does better look like? What's explicitly not in scope? Then push where it's soft — "slow", "confusing" and "better UX" aren't outcomes, and a problem without a number in it usually hasn't been looked at yet. Stop when you could hand it to someone who's never heard of it. ## The file ```markdown # Intent: drafts survive a refresh Raised by: T. Okafor (support). Status: draft. ## Problem Long-form posts are lost when the editor tab reloads. 14 support tickets this quarter; ev