fix(mobile): stop crashing on launch before the shell snapshot arrives

[juliusmarminge]

Build 1.1.1 (56) still crashed on every cold open even with #11486 in it. TestFlight strips the JS error, but expo-updates writes it to the app container (Library/Application Support/expo-error.log), pulled off the device with devicectl:

Fatal error: Domain: RCTErrorDomain
Description: Unhandled JS Exception: TypeError: Cannot read property 'defaultModelSelection' of null
    at NewTaskFlowProvider (main.jsbundle:590585:51)
    at NavigationProvider …

resolveProjectSettings (added in #11176) reads the project aggregate's legacy defaultModelSelection/defaultThreadEnvMode while projectSettingsFolded is false. The mobile new-task flow (#10639) passes its selectedProject straight through, and that is null until the shell snapshot arrives — so the very first render of NewTaskFlowProvider throws, RN reports a fatal, and expo-updates' ErrorRecovery aborts the process ~1 s after launch. That is the same native trace as the toReversed crash, which is why it looked unfixed.

tsc never flagged it: apps/mobile/tsconfig.json extends Expo's base, not the repo's tsconfig.base.json, so noUncheckedIndexedAccess is off there and projectsForEnvironment[0] ?? null types as non-null.

Fix: accept null for the project argument and treat it like an absent project.

Verification

vp test run packages/shared/src/projectSettings.test.ts — new case fails on main with the exact production error and passes here:

# before
 FAIL  resolveProjectSettings > treats a null project like an absent one before the shell snapshot arrives
 TypeError: Cannot read properties of null (reading 'defaultModelSelection')
# after
 Tests  11 passed (11)

tsc --noEmit clean for packages/shared and apps/mobile. Not a visual change.

Claude Fable 5 via Claude Code.

---

Summary by CodeRabbit

• Bug Fixes
  • Improved project settings resolution when project data is unavailable or null.
  • Preserved default server settings until project information becomes available.
  • Added regression coverage for handling missing project and aggregate data.

[macroscopeapp]

Approvability

Verdict: Approved at 421ea38

Macroscope's review found this PR approvable — This is a narrowly scoped mobile crash fix that handles the already-observed null project state without changing valid project-settings resolution or product defaults. It includes focused regression coverage and has no deployment, schema, security, billing, or static-analysis implications.

^{You can add or adjust custom eligibility rules. Learn more.}

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Team

Run ID: 3afad2ca-07aa-4d10-9ee2-92c1ff47060a

📥 Commits

Reviewing files that changed from the base of the PR and between d7c71f9 and 421ea38.

📒 Files selected for processing (2)

• packages/shared/src/projectSettings.test.ts
• packages/shared/src/projectSettings.ts

Included review availability: 9 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 10 reviews per hour.

---

📝 Walkthrough

Walkthrough

resolveProjectSettings now accepts a nullable legacy project. Null project data is treated as absent. A regression test verifies that null project and aggregate inputs preserve default server settings.

Changes

Project settings resolution

Layer / File(s)	Summary
Nullable project resolution packages/shared/src/projectSettings.ts, packages/shared/src/projectSettings.test.ts	resolveProjectSettings accepts null for the legacy project argument and treats it like an absent project. The test verifies default settings for null project and aggregate inputs.

Priority: ➖ Normal

Estimated code review effort: 1 (Trivial) | ~5 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 421ea

The nullable project path is covered by regression testing, with no remaining merge-blocking risk identified.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the primary change: preventing a mobile launch crash before the shell snapshot arrives.
Description check	✅ Passed	The description explains the crash, root cause, fix, verification steps, and lack of UI changes. It does not use the template headings or include the checklist, but it provides the required substantiv…
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 2 files.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/project-settings-null-project

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.6 KiB	13.6 KiB	+78 B (+0.6%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	+1 B (+0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.6 KiB	+77 B (+1.2%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	57.1 KiB	+88 B (+0.2%)	66.4 KiB	✅
Codex	Live turn messages	8	10	+2 (+25.0%)	21	✅
Claude	Total thread wire	13.6 KiB	13.6 KiB	−35 B (−0.3%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	+4 B (+0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.5 KiB	6.5 KiB	−39 B (−0.6%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	−44 B (−0.1%)	66.4 KiB	✅
Claude	Live turn messages	9	8	−1 (−11.1%)	21	✅

Baseline: d7c71f9 · PR result: 421ea38 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 113.9 KiB
• Claude decoded thread snapshot: 114.6 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.