Draft: two-way teleport with native CLI sessions

[baanish]

Draft. Do not merge. Exploration of two-way teleport. Opened as one PR because GitHub cannot stack cross-fork bases on this repo.

What Changed

T3 can list, import, and export native CLI sessions for Codex, Claude, Grok, and OpenCode.
Native files stay native. Import copies a CLI session onto a T3 thread. Export writes the thread back. Presence (t3 vs native) swaps Import/Export so both sides are not writers at once.
Codex export writes session_meta the CLI can resume. OpenCode imports text turns only, does not treat live opencode.db as a foreign lock, and does not match parent cwds into sibling codex / claude / grok folders. Stop settles Working when the turn is aborted, when there is no live session, or when interrupt fails.

Why

Two-way teleport needs a shared engine, UI, and one adapter per CLI. This PR is that whole stack in one diff because stacked bases have to exist on pingdotgg/t3code.

UI Changes

Import lives on the new-thread / sessions view. Export lives in the chat header. Composer send is disabled while the thread is in the native CLI.

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for any UI changes
• I included a video for animation/interaction changes

Note

Add two-way teleport between app threads and native CLI sessions

• Introduces a full teleport system allowing threads to be exported to native CLI sessions (Claude, Codex, Grok, OpenCode) and imported back, with message history synchronization via new thread.history.replace and thread.teleport.set orchestration commands.
• Adds TeleportService with list/import/export session RPCs (teleport.listSessions, teleport.importSessions, teleport.exportSession) wired into the WS layer with scope-based authorization.
• Adds per-provider JSONL/directory format adapters with schema-version gating, file-lock detection, and atomic writes; a registry pattern enables runtime format registration.
• Extends the command palette and chat UI with session browsing, import flows, a TeleportOutButton in the chat header, and composer/send disabling when a thread is teleported-out.
• Adds DB migration 041 adding teleport_json to projection_threads with backfill from existing runtime telemetry.
• Risk: thread.turn.start now fails with OrchestrationCommandInvariantError if the thread's teleport presence is native, blocking turn creation on teleported-out threads.

📊 Macroscope summarized 18ed3c0. 52 files reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0209abea-ab9a-4e89-8112-0400ee85f9af

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[macroscopeapp]

Two small consistency findings in the new teleport UI. Everything else in the changed web files (command palette views, composer disabled/placeholder handling, ChatView toast) follows the existing primitives and layout conventions.

^{Posted via Macroscope — UI Consistency}

[macroscopeapp]

Effect service conventions review of the new TeleportService (plus its contract errors). Four convention violations found; details inline.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

One finding: the new header control's disabled state suppresses the tooltip that explains why it is unavailable.

^{Posted via Macroscope — UI Consistency}

[macroscopeapp]

One consistency finding on the new header control; everything else (palette icon classes, variant="outline" size="icon-xs" matching the sibling header triggers, composer disabled/placeholder wiring) lines up with the existing patterns.

^{Posted via Macroscope — UI Consistency}

[macroscopeapp]

Reviewed the new teleport service and its call sites against the Effect service conventions. One finding: the native format adapters are wired through a mutable module-global registry populated by import side effects rather than through an Effect service/layer.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

One finding: the new teleport lock probe spawns an external command through a module-global execFile instead of acquiring the existing ProcessRunner service, so that dependency is not represented in the Effect environment. Everything else in the new TeleportService / TeleportFormatRegistry modules follows the canonical tag → make → layer shape, acquires its dependencies with yield* Foo.Foo, and uses Effect.catchTags for known tagged failures.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Two error-modeling convention issues in the new teleport code: message stored as data when it is a fixed template over existing attributes, and a helper whose only behavior is constructing that error.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Reviewed the new teleport service against the Effect service conventions. Service shape (Context.Service with inline interface, exported make/layer), dependency acquisition via yield*, and Effect.catchTags usage all look right now. Three remaining error-modelling issues below.

^{Posted via Macroscope — Effect Service Conventions}