[juliusmarminge]

Problem

A detached thread could not switch providers when the outgoing provider had never returned a native thread reference, even though the server could hand off its saved history.

Change

Allow the existing portable handoff once there is no live session or active run. Keep negotiated session capabilities and the active-run guard authoritative. Web, desktop and mobile consume the shared helper; no server or adapter behavior changes.

Scope and approval

This is a focused fix to the existing provider-switch workflow. Retains the work by @FabioLukas123 in #19687, with the pinned main revision merged and original commits preserved. It adds no providers, model IDs or wire fields.

Verification

• The regression is one changed assertion in an existing test. It fails without the fix; ten redundant PR-added cases were removed without deleting main tests.
• 22 focused shared/mobile tests pass. Existing tests still reject active runs and sessions that cannot hand off; a guard-removal probe makes them fail.
• 11 server planner tests and five existing portable-text-handoff integration scenarios pass. An independent review also passed 170 targeted shared/mobile/server/web tests. Those existing integration scenarios cover the transfer path; the live fixture below covers the missing-native-reference case.
• Targeted lint, formatting and client-runtime typecheck pass. Independent Opus and Sol full-diff reviews found no blockers.
• Root compared 952921f39a with 798ab31708 in the actual web app using a synthetic detached thread with no native reference. The alternate provider is disabled before and enabled after. A session denying handoff and an active run each keep it disabled.

Before: alternate provider disabled	After: alternate provider enabled

Control: live session denies handoff	Control: active run blocks switching

Root selected Fixture Handoff Target and sent through the real app pipeline to a copied repository Codex mock peer. The transport record contains thread creation, the exact prior-history marker T3_19687_CONTEXT_MARKER in thread/inject_items, and a completed subsequent turn. This proves history reached the target transport; it makes no claim about external model comprehension.

No external model or network call was made by the mock. Native desktop/mobile shells were not independently exercised. An identical temporary startup-recovery bypass preserved synthetic control rows on both variants; it was removed, the isolated database/settings restored, and the test server stopped.

Current-head CI runs after this update; earlier checks cover the original head only.

Implementation: Claude Fable 5.1. Test assessment and independent review: Claude Opus 5.5. Independent review and coordination: GPT-6.1 Sol, through Claude Code and Codex in T3 Code.

Takeover

Credit to @FabioLukas123 for identifying the detached-session regression in #19687.
Rebuilt on the pinned main revision: remove only the native-reference prerequisite; keep the source provider-thread association so saved history is handed off rather than silently omitted.
The original description above is preserved as historical context; verification below describes this replacement at 271e7b6f11.

Evidence

Before 952921f39a, after 271e7b6f11. Baseline/delivery matrix, completed controls/routes matrix.

Behavior	Proof and result
Detached completed source, null native reference	Before blocks target; after selects target. Seeded visual pairs at 1440/390, light/dark; raw and highlighted diffs in artifact index.
Saved history reaches target	Real-app sequence: completed synthetic source → actual detach → stopped-server null-native-ref fixture → target turn receives prior sentinel and completes.
Attached session true/false	True selects target; false stays blocked. False is an explicitly injected WebSocket projection checked in effective client state, not real adapter negotiation. Control recordings and projections.
Active work and missing source association	Event-backed held preparing run with no session stays blocked; missing source association stays blocked. Running status also has focused helper coverage.
Alternate picker routes	Keyboard opening, provider navigation, numbered jump, Favorites/search actually select target. Numbered jump uses supported opt-in web Alt+1 binding.
Native surfaces	Mobile unverified: Device access disabled despite installed Android tooling. Electron shell/default desktop shortcuts unverified; responsive web is not native proof.

Verification log

• Tests: existing shared file stays 21 cases; 4 mobile wrapper cases. No-native-ref regression failed before fix; missing-source guard regression failed against the overbroad prototype. Final head: 61 tests pass across shared workflow, mobile wrapper, server transition policy, and provider-switch integration files.
• Targeted formatting, lint, client-runtime typecheck, and diff checks pass. No repo-wide checks run.
• Review round 1: one FIX_FIRST and one MERGE. Retained source-association guard; rejected cancelled-only restriction because server intentionally accepts no-context switches. Round 2: independent GPT-6.1 Sol and Claude Opus 5.5 both MERGE.
• Proof audit: pass 2 checked 121 hashes, 44 images, five recordings and independently rebuilt the picker. Supplemental missing rows were captured; pass 3: SUFFICIENT, with audited row matrix.
• Server already accepts detached portable handoff without a native reference; target capability checks remain authoritative. No server, contract, auth, permission, mode, sandbox or MCP changes.
• All data synthetic, isolated dev homes, headed Chromium/Xvfb and ffmpeg x11grab. Original bundle, supplemental scripts/fixtures/bundle. Owned capture and audit processes stopped; temporary worktree/state cleanup follows publication.
• CI: pending on replacement PR; dispatcher owns follow-up. No merge performed.
• Friction: Device access disabled; incorrect initial cancelled-only fixture assumption corrected; source-reference probe initially copied an older whole file; subscription cache and picker automation required fixture corrections. Capture friction details.
• Models/harness: Claude Opus 5.5 tests/implementation/review; GPT-6.1 Sol coordination/review/capture/proof audit, through T3 Code Claude and Codex harnesses.

[macroscopeapp]

Approvability

Verdict: Approved at 271e7b6

Macroscope's review found this PR approvable — This focused client-runtime fix removes only an obsolete native-reference prerequisite for completed detached threads while retaining source-thread, session-capability, and active-run guards. It adds regression coverage without changing schemas, server adapters, product defaults, or static-analysis configuration.

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

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: takeover/pr-19687
• File: .coderabbit.config.ts
• Commit: 271e7b6
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Team
• Run ID: ca0fc9a4-9a5b-4cd2-aad9-5cc0aeec80de

Commits

Reviewing files that changed from the base of the PR and between 453bf01 and 271e7b6.

Files selected for processing: 2

• packages/client-runtime/src/state/threadWorkflows.test.ts
• packages/client-runtime/src/state/threadWorkflows.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 1 remain after this review.

Walkthrough

The provider handoff eligibility check no longer requires a native thread reference on the active provider thread. Tests cover matching provider-thread conditions and ready-session capability checks.

Changes: Provider handoff eligibility

Layer / File(s)	Summary
Update handoff eligibility and tests packages/client-runtime/src/state/threadWorkflows.ts, packages/client-runtime/src/state/threadWorkflows.test.ts	A matching active provider thread can qualify for handoff without a native thread reference. Tests cover the absence of provider threads and verify that a ready attached session qualifies only when it supports provider switching via handoff.

Priority: Normal

Estimated code review effort: 2 (Simple) | ~8 minutes

Change: Bug fix

Possibly related PRs

• pingdotgg/t3code#24579: Changes server delivery of portable context handoffs, which this PR enables the client to start when the source session is detached.
• pingdotgg/t3code#23971: Changes project and worktree handoffs around provider session replacement; this PR addresses client eligibility for a later provider handoff.
• pingdotgg/t3code#24545: Changes run finalization for successful worktree handoffs; this PR addresses client eligibility in the detached handoff flow.

Suggested reviewers: yash-singh1, macroscopeapp

Merge Risk: Minimal · up to 271e7

No concrete current-head defect or merge-blocking risk was established.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title "fix(clients): allow portable provider handoffs after detach" follows conventional commit format and matches the main change. The raw summary shows that threadSupportsProviderHandoff no long…
Description check		The description follows the repository template sections: Problem, Change, Scope and approval, and Verification. The Problem section states the bug and expected behavior. The Change section explains t…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Title check

Explanation

The title "fix(clients): allow portable provider handoffs after detach" follows conventional commit format and matches the main change. The raw summary shows that threadSupportsProviderHandoff no longer requires a non-null nativeThreadRef on the active provider thread, and the PR objectives confirm this removes the native-thread-reference prerequisite for switching providers on a detached thread. The title names the affected area (clients), the behavior (portable provider handoffs), and the condition (after detach). It is concise and specific, and it does not use vague terms.

Full details: Description check

Explanation

The description follows the repository template sections: Problem, Change, Scope and approval, and Verification. The Problem section states the bug and expected behavior. The Change section explains the fix and the components it touches. The Scope and approval section links the original work in #19687, explains the takeover, and states why the change is focused. The Verification section gives specific tests, results, and manual checks, and it notes what was not checked, including mobile and native desktop shells. The description is long and includes some process detail, such as model names and review rounds, that does not affect the template requirements. The required information is present. The Scope and approval section does not link a maintainer approval comment for the direction, which the template asks for when the change is not a very small obvious fix. This is a gap, but the section otherwise addresses scope.

Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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