Lamina controls and menus
Headless Product Design skill for AI coding agents | Design how it works, verify what you ship.
npx -y skills add aryaniyaps/lamina --skill lamina-controls-and-menusAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its author says it does
Copied from the file, not written here
Actions and reversibility in contracts — primary/destructive actions, undo policy, confirmations. Use when destructive ops lack scenarios.
SKILL.md
2.3 KB, as published. Nobody here has run it
Controls and Actions (agent-native)
Specify action semantics in screen and workflow contracts — destructive, reversible, confirmation policy — not menu widget styling.
Contract encoding
Per action on surfaces[]:
id,label,destructive: true|falsereversible: true|false→ if true, confirmation usually unnecessaryconfirmation_requiredonly when undo impossible + scenario documentsundo_scope: step | session | none
Working-set actions on primary toolbar area in screen spec; full catalog in nav/menus.
Frameworks
- Ask forgiveness, not permission: undo replaces confirm dialogs for recoverable ops.
- Rich modeless feedback: status in view, not modal for normalcy.
- Preview/compare: before irreversible-feeling changes — note in workflow step.
Design checklists
- Destructive ops have
scenarios[]+ recovery or undo policy. - Daily working-set actions ≤1-2 clicks from working screen.
- Confirm only when invariant violation risk and undo absent.
- Selection + manipulation pattern documented for sovereign apps.
- Platform conventions noted in
implement.md(web vs desktop) — Lamina stays unopinionated on library.
Verify checks
- Actor walk: destructive op matches contract confirmation/undo behavior.
- Accidental click recovery — undo or cancel path exists per contract.
Anti-patterns
- Confirm dialog for undoable delete.
- Modal error for recoverable validation.
- All actions equal weight on crowded screen spec.
- Hidden working-set functions three levels deep.