[maria-rcks]

people whose nightly desktop app cannot load cannot reach its update controls. npx t3@nightly unfuck starts a guided recovery session through the existing triage launcher, with claude/codex selection, machine context, and a recovery playbook that checks main for current instructions.

it distinguishes desktop replacement from cli/server updates, checks for a published fixed build, preserves the installed channel and user data, and asks before replacing software or restarting work. without an installed agent it prints manual recovery guidance and saves the prompt for another agent. existing triage behavior is preserved.

verified:

• all 8 focused tests pass, including both commands' no-agent fallback with an existing database/settings left unchanged and the bundled playbook matching its canonical copy.
• real unfuck --help and no-agent command runs pass without starting a server or creating a database.
• real unfuck --agent codex --model gpt-6.1-sol launch: codex read the prompt and context, asked desktop versus cli/server and affected machine, checked published releases, and used the bundled copy when the not-yet-merged online playbook returned 404. no installed app was changed.

scoped checks and remaining ci jobs are running. installer replacement and windows/linux interactive sessions remain unverified. this is a guided recovery command, not an automatic desktop installer.

model: unknown-model. harness: codex.

[coderabbitai]

Warning

Review limit reached

Only developers with an assigned seat can use this organization's usage-based review budget, and seats here are assigned manually. Ask an admin to assign a seat, or change the review continuation mode in Billing.

Next included review available in 11 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 10 included reviews currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: feat/nightly-recovery
• File: .coderabbit.config.ts
• Commit: 2678e20
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Advanced
• Run ID: fcf872c5-64e8-404b-bd6b-68ad0c9d441d

Commits

Reviewing files that changed from the base of the PR and between 6a3a460 and 2678e20.

Files selected for processing: 8

• .github/recovery/PLAYBOOK.md
• apps/server/src/binCli.ts
• apps/server/src/cli/recoveryPrompt.ts
• apps/server/src/cli/triage.test.ts
• apps/server/src/cli/triage.ts
• apps/server/src/cli/triagePrompt.test.ts
• apps/server/src/cli/triagePrompt.ts
• docs/user/updating.md

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

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

[macroscopeapp]

Approvability

Verdict: Needs Human Approval

Macroscope's review found this PR requires human approval — This PR adds a new t3 unfuck recovery workflow that launches an external coding agent and guides users through potentially consequential app or server updates, replacements, and restarts. Its explicit opt-in and preservation guidance reduce risk, but the new user-facing capability and operational scope merit human review.

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

[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	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.9 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	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	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 6a3a460 · PR result: 2678e20 · Source CI: cancelled

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.