Loading request...
The current design issue workflow puts all UX questions on the developer. The user proposes that when the AI picks up a design issue, it should propose a resolution for each open UX question with a brief rationale, rather than just asking.
## Problem The current `[2]` design issue workflow puts all UX questions back on the developer to answer. In practice, the developer may not have strong opinions on many of them and would prefer to just accept a sensible AI proposal rather than think through every question independently. ## Proposed improvement **When the AI picks up a `[2]` design issue, it should:** 1. **Propose a resolution for each open UX question** — not just ask. Frame it as a recommendation with a brief rationale. The developer can accept by saying nothing, or comment to push back. Example format: > **Date range selector** — I recommend preset buttons (1y / 2y / 5y / 10y / 20y / Max) with Max as default. Rationale: gives the full historical context on load (which is the core value of the app) while letting the user zoom into shorter windows easily. Does this work? 2. **Classify scope questions correctly** — when a new feature idea surfaces during design, the AI should recognise it has three options