Why now
The iOS app compiles cleanly for the iOS Simulator but has never run on physical hardware — the Simulator has no camera. This is already recorded in two places:
PROJECT_STATUS.md (2026-08-11): "App Store Connect record creation and physical camera/ReplayKit verification remain separate manual gates."
foundry-evidence.json: device.state: blocked — "No physical device in CI; camera + ReplayKit need a device."
#20 (App Store Connect record + TestFlight upload) was deferred by the owner on 2026-08-13 specifically until this verification is complete. Since #20 was the repository's only open issue, the gate it waits on was untracked — #20 pointed at work that had no home. This issue is that home.
It is also Motion's reopen condition in the fleet portfolio: "Reopen for focused physical-device playtesting of one genuinely fun control loop."
Acceptance
The first seven are pass/fail engineering checks. The last one is the actual decision. README.md and PROJECT_STATUS.md both state that the central product risk is control feel, not infrastructure — which is why v1 deliberately strips out every server and network hop. Nothing currently asks that question.
Out of scope
Done when
Recorded evidence exists (FPS, input latency, one saved recording) alongside a written verdict on control feel, and then either:
Why now
The iOS app compiles cleanly for the iOS Simulator but has never run on physical hardware — the Simulator has no camera. This is already recorded in two places:
PROJECT_STATUS.md(2026-08-11): "App Store Connect record creation and physical camera/ReplayKit verification remain separate manual gates."foundry-evidence.json:device.state: blocked— "No physical device in CI; camera + ReplayKit need a device."#20 (App Store Connect record + TestFlight upload) was deferred by the owner on 2026-08-13 specifically until this verification is complete. Since #20 was the repository's only open issue, the gate it waits on was untracked — #20 pointed at work that had no home. This issue is that home.
It is also Motion's reopen condition in the fleet portfolio: "Reopen for focused physical-device playtesting of one genuinely fun control loop."
Acceptance
README.md.WKWebViewgame content rather than a black frame. The second risk called out inREADME.md.webgame/path works with no Vite dev server and no Mac on the network.The first seven are pass/fail engineering checks. The last one is the actual decision.
README.mdandPROJECT_STATUS.mdboth state that the central product risk is control feel, not infrastructure — which is why v1 deliberately strips out every server and network hop. Nothing currently asks that question.Out of scope
Done when
Recorded evidence exists (FPS, input latency, one saved recording) alongside a written verdict on control feel, and then either: