lamina-discoverability

# Discoverability (agent-native) Specify **what actors can perceive and do** on each screen — signifiers, primary actions, disabled reasons — so verify catches gulfs of execution and evaluation. ## Contract encoding Per `surfaces[]`: - `primary_actions[]` — visible, labeled - `secondary_actions[]` — deferred or progressive disclosure - `disabled_when` — guards linking to scenarios - `feedback` — what actor sees after each action (ties to `feedback-and-status`) ## Diagnostic (verify) | Gulf | Symptom in actor walk | Fix in contract | |------|----------------------|-----------------| | Execution | Can't find how to act | Add signifier / promote primary action | | Evaluation | Can't tell if action worked | Add feedback scenario + screen state | ## Design checklists 1. Primary workflow actions visible without hunt. 2. Constraints prevent wrong actions; don't rely on warnings alone. 3. Mapping: control position relates to effect when spatial layout matters. 4. Consistent labels for same operation across screens. 5. Actor self-blame in simulation → design failure finding. ## Verify checks - Primary actor completes workflow without instructions document. - Walkthrough: mystery meat / hidden gestures flagged. ## Anti-patterns - Invisible affordances (gestures with no signifier). - Hidden modes changing control meaning. - Identical unlabeled controls. - Assuming actors learn internal jargon. ## Related - [Feedback And Status](../lamina-feedback-and-status/SKILL.md) - [Onb...