[pseudoseed]

Merges upstream pingdotgg/t3code main at dd14848153 (133 commits) into the fork and bumps PseudoCode to 0.0.77.

Conflicts

12 files conflicted. Each kept both sides:

• Version files (4 package.json): kept the fork's version line, then bumped to 0.0.77 in its own commit. Upstream is on 0.1.5.
• AddUsageLimitSourceDialog: upstream moved sourceIdFromUrl into AddUsageLimitSourceDialog.logic.ts and made it avoid ID collisions. The fork's kind (CLIProxyAPI hub or usage dashboard) now passes into that function, which matches saved sources on kind and URL. The upstream test passes "cliproxy".
• ProviderEventIngestor: the fork's usage lock and upstream's usage-limit refresh tap both run.
• ChatView: the fork's Issues surface callback and upstream's thread-find callback are both kept.
• ThreadLineageWorkflowRow: upstream's Spinner replaces LoaderCircleIcon; the fork's EnvironmentId/ThreadId imports stay.
• SettingsPanels: both imports kept.
• serverRuntimeStartup.autoPull.test: the mock keeps the fork's streamDomainEvents/getShellSnapshot, which DirectPush needs at startup, plus upstream's countThreads.
• terminalUiStateStore.test: both tests kept.
• agentContextBudget.json: set to the measured values. mcpToolListJsonBytes goes from 48004 to 48104 (+100) because of upstream's merged MCP tools. The Agent context budget check needs the agent-context-increase label, which does not exist on this repo yet.

Fork changes survived

fork-delta-check.py -v (run on the merge commit, before this PR merges): 0 DROPPED. It listed 1 REWRITTEN file, AddUsageLimitSourceDialog.tsx (9 lines); its kind-aware ID logic now lives in the .logic.ts file described above. The fork's 2 dropped comments there are intentional.

Tests

• vp run typecheck: 36/36 packages pass.
• vp run lint: pass (warnings only).
• run-suites.sh: 31 of 32 packages pass. That includes desktop 875, mobile 3080, server 5484, client-runtime 1496, shared 1113, and web 5000 of 5001.
• 1 failure: GuestPage.browser.test.tsx "follows the end as replies stream…" (expected 2074 to be less than 2). It fails the same way on a clean upstream/main worktree on this Mac (1 failed, 2 passed). It is the known Mac-only failure in the sync skill.
• check:changed also failed 2 desktop Linux icon tests. They write file names that differ only by case, which this Mac's case-insensitive disk cannot hold. run-suites.sh runs desktop on a case-sensitive volume, where they pass.

Live check

Ran against a worktree backend with copied real data (35 threads across 2 projects).

• Web:
  • App title and copy say PseudoCode.
  • Settings > Providers > Add hub shows both fork kinds. Adding a usage dashboard at http://127.0.0.1:8787 stored the ID aiusage-127.0.0.1-8787 with kind aiusage.
  • A Replay (dev) showcase turn streamed, asked for approval, and completed.
  • The right panel's Issues surface lists issues.
  • The terminal drawer opens.
  • The server log had 0 errors.
• iOS (iPhone 18 Pro simulator, dev client):
  • Paired, and the thread list loaded.
  • The replay thread renders in full.
  • Usage > Limits shows the dashboard source ("dashboard could not be reached", as expected with nothing listening).
  • Environment detail reports Version 0.0.77.
• Not from this merge: the web usage source row is hardcoded to say "CLI Proxy" for every kind. Fork main does the same.

Merge with a merge commit, not a squash.

🤖 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	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.3 KiB	1.3 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.6 KiB	21.6 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.1 KiB	+6 B (+0.1%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.3 KiB	1.3 KiB	+6 B (+0.5%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 1f5c51a · PR result: 25fc017 · 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: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

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