lamina-feedback-and-status

# Feedback and Status (agent-native) Every mutating operation declares **what the actor sees during and after** — in workflow steps and `scenarios[]`. ## Contract encoding Per workflow step: - `feedback`: immediate | progress | async_poll | silent_ok - `visible_state`: what changes on screen (entity state, badge, banner) - Async: `status: processing` scenario until terminal state **Modeless feedback** preferred over modals for routine status. ## Frameworks - **Visibility of system state**: actor always knows where they are in multi-step flows. - **Forcing functions**: only for safety-critical irreversible ops — document in scenario. - **Feedforward**: preview consequence before commit (destructive ops). ## Design checklists 1. No dialog reports normal completion. 2. Disabled controls explain why (link to scenario or inline hint). 3. Progress for ops > ~1s perceived wait. 4. Optimistic UI only when `consistency-guarantees` allow + rollback scenario exists. 5. Group related status in same screen region. ## Verify checks - Actor walk: each mutating step produces expected visible feedback. - Async probe: refresh during processing — stale state handled per invariant. - Double-click submit — idempotency feedback (`idempotency-concurrency`). ## Anti-patterns - Async silence — upload with no progress. - Disabled with no explanation. - "Are you sure?" without true irreversibility. - Modal blocking batch flow for recoverable status. ## Related - [Discoverability](../lamin...