apple-device-validation
Build, install, run, render, and validate Apple apps on the correct destination. Use for simulator/device evidence, screenshots, hardware, paired devices, and infrastructure triage.
SKILL.mdskills/apple-device-validation/SKILL.md
GitHubApple runtime and device validation
Inputs
Define the exact claim, authorized actions, source state, project/workspace, scheme, configuration, bundle, destination, required app state, and available tools. Discover current destinations instead of remembering device identifiers.
Rules
- PROOF-001 — Match evidence to claim. Source, compile, tests, install, launch, rendering, interaction, hardware, and cross-device proof are distinct.
- PROOF-002 — Respect action boundaries. Installation does not authorize uninstall, data reset, account changes, or unrelated source edits. Advance only as the requested validation requires.
- PROOF-003 — Reproduce the environment. Record exact source, toolchain, destination/runtime, fixture state, commands, exit status, and relevant artifacts.
- PROOF-004 — Classify infrastructure first. Separate unavailable runtimes, pairing, service health, provisioning, installation, and application defects before changing code.
- PROOF-005 — Validate the current candidate. Repeat affected checks after changes; reconcile superseded receipts before reporting a blocker. Remove temporary harness state unless it is an intentional maintained test facility.
Workflow
- Set the minimum sufficient proof contract and inspect environment readiness.
- Run the narrow build/test prerequisite for the selected destination.
- Install, launch, interact, and capture evidence only within the authorized scope.
- Exercise relevant lifecycle, failure, persistence, and system-boundary transitions.
- Report the strongest supported claim and every blocked or unverified gate.
Verify
For layout, inspect rendered output at risky sizes and accessibility settings. For persistence, prove store readiness, mutation, and relaunch. For hardware or paired-device claims, test the real boundary rather than substituting simulator success.
Output
Return a compact evidence matrix with pass/fail/blocked/not-run, exact candidate and environment, artifacts, failure classification, and remaining device gates.
References
Read the playbook for decisions, failure cases, and source links. For a whole-app review or batch feedback, use the screen review workflow.
For debugger, simulator, and evidence integrity, read the focused reference.