epic: v2.0 coverage expansion — device fidelity, data inspection, perf, shipping, i18n

[conorluddy]

Goal

Expand the skill from 22 → ~39 scripts, covering the iOS DX workflows that indie devs and small teams hit daily but the skill ignores today.

Why now

1000+ users. Gap analysis (see PR #53 discussion) shows the next 80% of agent value is in inspecting and shipping, not just driving the app.

Clusters and child issues

A. Device fidelity (real-world state simulation)

• feat: location.py — GPS + GPX route playback #54 — location.py — GPS + GPX route playback
• feat: appearance.py — dark mode / dynamic type / locale #55 — appearance.py — dark mode / dynamic type / locale
• feat: simulate_event.py — memory warning / background / battery #56 — simulate_event.py — memory warning / background / battery

B. Data inspection

• feat: container.py — app sandbox + UserDefaults + Core Data #57 — container.py — app sandbox + UserDefaults + Core Data
• feat: Crash log parsing and analysis #36 — Crash log parsing and analysis (pre-existing)
• feat: entitlements.py — security & ATS audit #59 — entitlements.py — security & ATS audit

C. Performance & hangs

• feat: Instruments/profiling integration via xcrun xctrace #32 — Instruments / xctrace integration (pre-existing)
• feat: hang_watcher.py — os_log hang stream #60 — hang_watcher.py — os_log hang stream
• feat: hang_watcher session mode — start/stop, threshold, dedup, progressive disclosure #77 — hang_watcher.py session mode (start/stop, threshold, dedup, progressive disclosure)
• feat: launch_perf.py — cold/warm start measurement #61 — launch_perf.py — cold/warm start measurement
• feat: main_thread_sampler.py — sample a stuck PID #62 — main_thread_sampler.py — sample a stuck PID
• feat: metrickit_reader.py — MetricKit diagnostic payload reader #63 — metrickit_reader.py — MetricKit diagnostic payload reader

D. Shipping & build

• feat: deep_link.py — URL scheme + Universal Link tester #64 — deep_link.py — URL scheme + Universal Link tester
• feat: archive_and_upload.py — TestFlight pipeline #65 — archive_and_upload.py — TestFlight pipeline
• feat: build_settings.py — xcconfig / scheme / test plan inspect #66 — build_settings.py — xcconfig / scheme / test plan introspection
• feat: screenshot_matrix.py — multi-device App Store assets #67 — screenshot_matrix.py — multi-device App Store assets

E. Quality & i18n

• feat: localization_audit.py — string catalog gaps #68 — localization_audit.py — string catalog gaps
• spike: swiftui_previews.py — headless preview rendering #69 — swiftui_previews.py — SPIKE: headless preview rendering

Dependency notes

• B1 (feat: container.py — app sandbox + UserDefaults + Core Data #57) ships before C4 (feat: metrickit_reader.py — MetricKit diagnostic payload reader #63) — MetricKit reader walks the sandbox.
• A2 (feat: appearance.py — dark mode / dynamic type / locale #55) ships before D4 (feat: screenshot_matrix.py — multi-device App Store assets #67) — screenshot matrix uses theme/locale switching.
• C1 (feat: hang_watcher.py — os_log hang stream #60) ships before C3 (feat: main_thread_sampler.py — sample a stuck PID #62) — hang watcher triggers the sampler.

Definition of done

Epic closes when 14/17 children ship. The swiftui_previews spike (#69) may land as research-only without producing a script — its deliverable is a go/no-go decision plus findings. On close, cut a v2.0.0 release tag.

Provenance

Originates from the gap analysis in the PR #53 review conversation. See planning doc ~/.claude/plans/oh-interesting-lets-make-mossy-bee.md for the full design.