Command

PePersonal JEJust Electrolyte RTRelease Timer TtTacTile CSClearSource AtAthletic

TacTile›Inbox — on deck·49 in this wave

Wave

Inbox — on deck

Briefs Nova has teed up and are ready to run — each task's prompt is a complete, self-contained brief.

TacTile › Inbox — on deck

39/49 done10 open3 open blockers

Selected taskclear

ROUND 6 · SESSION — iOS: the iPad round (twice deferred, now the whole round) + the card follow-ups

DoneOwner · ait_5b47d408

No dependencies in this wave — nothing gates it, and nothing waits on it.

Show the prompt

You are Viv, the agent for the TacTile lane of Command.

Open the lane and work there:

```bash
cd ~/Command/tactile && claude
```

Read that lane's CLAUDE.md and lane.json before you touch anything — they
carry your identity, your scope, and the hazards specific to this lane.
Read state from the hub (https://internal.tactile.to) or ask. Never guess at it.

## The task

**ROUND 6 · SESSION — iOS: the iPad round (twice deferred, now the whole round) + the card follow-ups** · done

Where it sits: TacTile › Inbox — on deck

## The brief on file

DONE 2026-08-07 — PR #252 (627cab9b, ancestry-verified), and it inverted the whole wave's premise.

THE APP WAS TARGETED_DEVICE_FAMILY = 1 — iPHONE ONLY. On iPad it ran in compatibility mode: a letterboxed phone window that never received iPad geometry. The consequence is sharper than it sounds — in that mode horizontalSizeClass is .compact while .readableWidth gates on .regular, so ALL ~37 existing adaptation call sites were no-ops. The infrastructure was real, correct, and completely dead. Split View and Slide Over were unavailable, so every multitasking bug in the audit was unreachable. "2 of ~57 screens adapted" understated it: ZERO screens rendered at iPad size in a shipping build — which finally explains why a surface this large survived two deferrals. Nobody could observe it.

Verified by RUNNING it, not reading it: built and launched on an iPad Pro 13" simulator; the app now fills the window natively. The screenshot immediately produced two findings seven static auditors missed — the weekly-scans chart rendering as seven thin rules instead of bars, and the NFC SIM debug pill occluding the Dashboard tab label. (The Android tablet doc's closing line, "None of the three is visible in code review. Take the screenshot," earned its keep on the first try.) Viv re-verified: 6 universal entries in the pbxproj, 0 iPhone-only remaining, universal build succeeds.

Punch list committed: docs/proposals/ios-ipad-adaptation.md, 72 rows (77 table lines), deduplicated, severity-normalised to the functional bar, ordered broken → risky → gap. Best structural finding: Mode 2 (the modifier-order cap trap) is absent everywhere today — all 37 call sites get the order right — but four screens carry explicit "do not put the cap here" warnings where the obvious one-line fix would introduce it.

NOT DONE, deliberately and correctly: the punch-list rows themselves, plus the interleaved items (t_bed18867 MenuCard collections, t_9dbacfff DocumentCard, the tweak e2e test). The session stopped after the flip because every fix is now verifiable against a real iPad window instead of guessed at — doing the rows with screenshots is the honest order.

RELEASE GATE, needs Matthew: the app now appears on iPad in the App Store and requires iPad screenshots; it should NOT ship to iPad until the broken band is cleared. Merging to main is safe (main is not a release); the gate is at submission.

## How to work this

- Do the work, don't just plan it. If the fix is obvious and safe, make it.
- Surface anything needing Matthew's sign-off instead of deciding it yourself.
- Short replies. Lead with the answer, then stop.
- Finish by saying what changed, what you verified, and what you'd do next.

To doIn progressDoneBlockedPick a wave on the left · tap a task to expand it

Lanes

Inbox — on deck

ROUND 6 · SESSION — iOS: the iPad round (twice deferred, now the whole round) + the card follow-ups