chore: sync upstream pingdotgg/t3code (57 commits) and fix macOS test failures

[donnes]

Upstream sync

Merges 57 commits from pingdotgg/t3code (through d06f0ff1 "fix(sqlite): retry failed statement preparations") into the fork as a single two-parent merge commit. Every fork feature survives; no fork change was dropped.

Conflicts resolved

• apps/server/src/persistence/Migrations.ts: upstream took migration 054 for auto-settle. The fork's Notes migration moved to 055_Notes.ts / registry entry 55.
• apps/web/src/components/chat/ChatHeader.tsx and ChatView.tsx: upstream rewrote the reserved-padding and transition classes in newer Tailwind syntax. Kept the fork's terminal-toggle spacing fix expressed in that syntax.
• apps/web/src/routes/__root.tsx: the auto-merge produced a duplicate isElectron import; kept upstream's alias form.

Test fixes

Four server tests failed on macOS before and after the sync:

• macOS temp dirs sit behind the /var -> /private/var symlink, so paths reported by git and by symlink-following skill discovery never matched the raw mkdtemp result. The git driver and skill-file test helpers now realpath the temp directory.
• The background-fetch env test set GIT_SSH, which a global core.sshCommand silently overrides on developer machines. It now sets GIT_SSH_COMMAND, which git gives precedence.

Verification

• vp i, no Effect catalog bump so .repos/ is unchanged
• Typecheck: contracts, shared, client-runtime, server, web, desktop, mobile all pass
• vp test run on the fork's 23 carry test files: 695 passed, 1 skipped, 0 failed
• Lint and vp fmt --check on every edited file: no errors

Merge method

Please merge this with "Create a merge commit". Squash or rebase would flatten the two-parent merge and make the next upstream sync re-conflict on everything. The current ruleset only allows squash and rebase, so it needs the merge-commit method enabled first.

Model: Claude Fable 5.1, harness: Claude Code in T3 Code.

🤖 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	13.5 KiB	13.5 KiB	+22 B (+0.2%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	+10 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.5 KiB	+12 B (+0.2%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	56.3 KiB	56.3 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	10	10	0 (0.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	+12 B (+0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	+18 B (+0.2%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.4 KiB	6.4 KiB	−6 B (−0.1%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.0 KiB	57.0 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	9	9	0 (0.0%)	21	✅

Baseline: 1fe78bb · PR result: 73175e3 · 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: 114.0 KiB
• Claude decoded thread snapshot: 114.7 KiB

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