chore(sync): merge 288 upstream commits, keeping the fork's features

[pseudoseed]

Brings the fork up to upstream 20012eb (2026-10-02): threads without a project, projects from a name, per-thread auto-settle, Working section, restart agent session, one-click provider updates, Codex 0.159 and ChatGPT managed auth, Claude Sonnet 5.5, and Expo SDK 58 / React Native 0.88 RC.

How the fork's work was kept:

• 71 conflicts resolved keeping both sides; every fork-added line since the merge base was checked against the merged tree. Phone project cards and iPad sidebar projects are rebuilt on upstream's new thread list.
• Upstream's new migration 54 runs as 58 after the fork's own; verified against a copy of real data.
• Expo 58 drops the fork's permission-registry lock and reload view fix; both carried as patches/expo-modules-core@58.0.9.patch with their regression tests in CI.
• Workflows take upstream's structure on GitHub-hosted runners.

Follow-up commits on this branch:

• Fork panels (Issues, mailbox, usage dial, MCP settings) follow upstream's now-mandatory shadcn lint rules.
• All user-facing "T3 Code" / "T3 Connect" text says PseudoCode / Cloud Connect. Data directories, package names, and values sent to OpenAI are unchanged.
• iOS builds on RN 0.88 with the voice module (Podfile hook for T3Voice's module map).
• Widget project icons and the update install prompt load on RN 0.88 (dynamic import("react-native") throws on the removed InteractionManager).

Verified: typecheck in 7 packages, ~6,800 tests, lint, fmt, knip, mobile native lint. iOS dev build on the iPhone 17 Pro simulator against copied real data: pairing, project cards, thread view, terminal, settings, environment detail, 3 background/resume cycles, no crashes. Not verified: Android, dictation, home-screen widget rendering, update prompt at runtime.

Merge with a merge commit so the upstream sync reverts as one unit.

Claude Opus 5.5 (1M context) via PseudoCode.

🤖 Generated with Claude Code

[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	13.9 KiB	14.0 KiB	+56 B (+0.4%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.3 KiB	7.3 KiB	+16 B (+0.2%)	7.4 KiB	✅
Codex	Live turn WebSocket wire	6.6 KiB	6.7 KiB	+40 B (+0.6%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.6 KiB	57.6 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	9	9	0 (0.0%)	21	✅
Claude	Total thread wire	14.0 KiB	14.0 KiB	+25 B (+0.2%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.3 KiB	7.3 KiB	+21 B (+0.3%)	7.4 KiB	✅
Claude	Live turn WebSocket wire	6.7 KiB	6.7 KiB	+4 B (+0.1%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	58.4 KiB	58.4 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	9	9	0 (0.0%)	21	✅

Baseline: bd962bb · PR result: 0136972 · 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: 116.3 KiB
• Claude decoded thread snapshot: 117.0 KiB

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