iOS Simulator Automation

Reuse compatible infrastructure, but bind observations to the runtime that produced them.

Choose the workflow

For a healthy one-agent check with no shared evidence or configuration matrix, use:

preflight -> open -> snapshot -> act -> re-snapshot -> assert -> cleanup

Use the runtime ledger for feature delivery, cross-agent or reusable evidence, multi-criterion matrices, or a shared retry budget. Authentication alone does not require the ledger.

Route and load

• Read the current agent-device skill for app-owned UI, assertions, app diagnostics, and React Native inspection.
• Read the current serve-sim skill for SpringBoard, system UI, hardware input, or system gestures.
• Read only the applicable recipe: Expo, authentication/backend, or system/accessibility.

Upstream skills and CLI help own command syntax. Stay with the tool that owns the current surface.