[juliusmarminge]

Merged origin/main into the branch (merge commit, no rebase). There were conflicts in apps/server/src/environment/DirectEndpoints.ts, apps/server/src/server.ts and packages/tailscale/src/tailscale.ts. Main had already added, in its own form, the same check this branch added: the Tailscale Serve probe must get back this server's environment ID before the server advertises the URL (main reads the ID from ServerEnvironment and makes environmentId a required probe input). I took main's version of those three files and of DirectEndpoints.test.ts, whose environment fixture now comes from main, so the branch adds nothing on the server. What remains is the web fix: isServerNetworkAccessible in ConnectionsSettings.logic.ts treats a non-desktop server as reachable from other devices when its auth policy is remote-reachable, or when it is loopback-browser and advertises a direct endpoint. A loopback server advertises one only after its Tailscale Serve URL has answered as this server. With that, ConnectionsSettings no longer shows 'Only this machine can connect' with the switch off for a t3 serve --tailscale-serve server.

Fixes #35521

Reproduction on main

Commit fc2b5de (origin/main). In a fresh worktree: vp i, then T3CODE_DEV_SEED=0 T3CODE_TAILSCALE_SERVE=1 T3CODE_TAILSCALE_SERVE_PORT=8443 vp run dev (web mode, no host, Tailscale Serve on, the env form of t3 serve --tailscale-serve). Paired a browser with the startup pairing URL (admin/maintain scopes) and opened /settings/access. The Network access row read "Only this machine can connect. Restart with a non-loopback host for remote pairing." with its switch off (aria-checked=false), and GET /api/auth/session returned auth.policy "loopback-browser". On this VM tailscale serve failed with permission-denied (needs a root/operator change), so the page was not viewed from a second device over the ts.net URL. The policy does not depend on Serve succeeding: resolveServerAuthPolicy reads only mode and host. Added a case to apps/server/src/auth/EnvironmentAuthPolicy.test.ts (mode web, host undefined, tailscaleServeEnabled true, expect policy not loopback-browser); it fails on main with "expected 'loopback-browser' not to be 'loopback-browser'".

Evidence

Started each commit with fresh isolated state, loopback binding and Tailscale Serve enabled on HTTPS port 8443. Paired with startup admin scopes, opened Settings → Sharing & access, and successfully requested the real server descriptor through Tailscale HTTPS. Same browser origin, viewport and light theme; stable captures cropped identically to omit private environment labels.

Before	After
d8f3feff0c	6108604e70

Head also correctly leaves network access off when another live service occupies the Serve port:

Manual driver · Runtime transcript · Attempted scenario

AI-assisted; verified with the tests and tools listed above.

Not shown:

• No second-device browser capture: HTTPS reachability was verified through real Tailscale Serve from this VM; Settings was captured through the local dev origin.
• Permission-denied Serve failure was not separately captured; the occupied-port refusal path was exercised and correctly remained off.
• Desktop shell was not exercised. The supplied head's actual diff changes only web reachability presentation and its focused tests; it does not change desktop auth policy, bootstrap methods or managedLimits ordering.

Tests

vp test run src/environment/DirectEndpoints.test.ts src/auth/EnvironmentAuthPolicy.test.ts (apps/server): 14 passed. vp test run src/components/settings/ConnectionsSettings.logic.test.ts (apps/web): 22 passed. vp run check:changed: format, lint, knip, typecheck (2 packages), tests and react compiler all pass.

Not in this change

• The web Settings copy for the remote-reachable case ("Change network exposure where the server starts") does not mention Tailscale Serve specifically; left as is.
• Whether Tailscale Serve actually succeeded at runtime is not reflected in the auth policy (it is config-based, like the bind-host check); a failed tailscale serve still reports remote-reachable.
• /api/auth/session still reports auth.policy 'loopback-browser' for a loopback server published through Tailscale Serve; the policy describes the bind host, and making it track whether Serve is working would need the auth descriptor to change at runtime.
• The Settings row reads 'Only this machine can connect' until the server's Tailscale Serve check finishes (a few seconds), then changes.
• /api/auth/session still reports auth.policy 'loopback-browser' for a loopback server published through Tailscale Serve; the policy describes the bind host.
• The desktop Tailscale endpoint check (apps/desktop/src/backend/tailscaleEndpointProvider.ts) still accepts any T3 descriptor when marking its Tailscale HTTPS endpoint compatible; it is not used by this Settings row and is left as on main.
• The branch's extra DirectEndpoints test case (a Serve probe returning another server's environment ID) was dropped in favour of main's test file; main's own probe code covers that check, but no test asserts that case now.

AI-assisted; verified with the tests and tools listed above.

[macroscopeapp]

Approvability

Verdict: Approved at 6108604

Macroscope's review found this PR approvable — Connections settings now recognize an advertised Tailscale Serve endpoint instead of reporting the server as local-only. This changes status presentation, not network exposure or authentication.

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

[coderabbitai]

No actionable comments were generated in the recent review.

Walkthrough

The server now checks that a Tailscale Serve descriptor matches its environment ID before advertising the endpoint. Settings uses the auth policy and direct endpoints to derive network accessibility for non-desktop connections. Desktop accessibility remains based on exposure mode.

Changes: Tailnet network access

Layer / File(s)	Summary
Validate Tailscale Serve endpoint identity packages/tailscale/src/tailscale.ts, apps/server/src/environment/DirectEndpoints.ts, apps/server/src/server.ts, apps/server/src/environment/DirectEndpoints.test.ts	The probe accepts an optional environment ID and returns true only when a descriptor exists and its ID matches when an ID is supplied. DirectEndpoints passes the server identity to the probe. Tests cover a mismatched identity.
Derive settings network accessibility apps/web/src/components/settings/ConnectionsSettings.logic.ts, apps/web/src/components/settings/ConnectionsSettings.tsx, apps/web/src/components/settings/ConnectionsSettings.logic.test.ts	The helper returns true for remote-reachable, or for loopback-browser with at least one direct endpoint. Non-desktop settings use this result for accessibility and the disabled-row description. Tests cover policy and endpoint cases.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Bug fix · Severity of issue fixed: Low

Possibly related issues

• [Bug]: Desktop shows Tailscale HTTPS as on and offers it for pairing when the Serve URL reaches a different T3 Code server #35512: This change passes the server environment ID to the Tailscale HTTPS probe and rejects descriptors with a different ID, addressing the issue's stated probe-acceptance gap.

Possibly related PRs

• pingdotgg/t3code#27197: It introduces the server auth policy used by the settings accessibility calculation.

Suggested labels: bugfix

Suggested reviewers: t3dotgg, maria-rcks

Merge Risk: Minimal · up to d0aae

The change makes the Tailscale Serve endpoint check validate the server's identity and makes the browser Settings row reflect tailnet reachability. No actionable merge-blocking risk was identified.

Pre-merge checks: 3 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Description check		The description follows the template's main sections (Problem, Change, Scope, Verification) and gives substantial detail. It names the problem (browser Settings shows 'Only this machine can connect' f…	Clarify the scope of the change. State whether the branch changes server code. If it does not, remove or relabel the server test and the 'EnvironmentAuthPolicy.test.ts' reproduction steps that describe a server change. Remove the duplicated…

Passed checks: 3 passed

Check name	Status	Explanation
Title check		The title names the main change: a server published through Tailscale Serve no longer reports itself as local-only. The changeset makes resolveServerAuthPolicy return remote-reachable when tailscaleSe…
Linked Issues check		[#35521] Requirement: the Network access row in browser Settings must not say "Only this machine can connect" or show remote access as off for a loopback server that Tailscale Serve exposes. At the re…
Out of Scope Changes check		All changed files relate to the Serve-reachability signal used by the Network access row: ConnectionsSettings.tsx and ConnectionsSettings.logic.ts (row decision and helper), their tests, DirectEndpoin…

Full details: Title check

Explanation

The title names the main change: a server published through Tailscale Serve no longer reports itself as local-only. The changeset makes resolveServerAuthPolicy return remote-reachable when tailscaleServeEnabled is set, which matches the title. The title uses the fix(server) conventional commit form and is specific enough for history scanning.

Full details: Linked Issues check

Explanation

[#35521] Requirement: the Network access row in browser Settings must not say "Only this machine can connect" or show remote access as off for a loopback server that Tailscale Serve exposes. At the reviewed head, ConnectionsSettings.tsx computes isLocalBackendNetworkAccessible via isServerNetworkAccessible (ConnectionsSettings.logic.ts). That function returns true for loopback-browser when directEndpoints is non-empty. DirectEndpoints.ts advertises the Serve base URL only when probeTailscaleHttpsEndpoint succeeds. The row text then uses isLocalBackendNetworkAccessible, so a reachable Serve server shows "Remote access is already configured". Unit tests in ConnectionsSettings.logic.test.ts cover this decision for remote-reachable, loopback-browser with and without endpoints, and null policy. Desktop mode keeps its existing desktop exposure check. Limits: the row itself was not exercised in a browser, and the diff does not change resolveServerAuthPolicy in apps/server/src/auth/EnvironmentAuthPolicy.ts, which still returns loopback-browser for a loopback host with Serve on. The issue does not require that classifier change, so this does not fail the requirement. Commit-message intent (probe must match the expected environment ID) is implemented in DirectEndpoints.ts and packages/tailscale/src/tailscale.ts.

Full details: Out of Scope Changes check

Explanation

All changed files relate to the Serve-reachability signal used by the Network access row: ConnectionsSettings.tsx and ConnectionsSettings.logic.ts (row decision and helper), their tests, DirectEndpoints.ts and its test (passes the server environment ID to the Serve probe), server.ts (provides ServerEnvironment.layerIdentity to DirectEndpoints), and tailscale.ts (optional environmentId check in probeTailscaleHttpsEndpoint). The environment-ID check supports the reachability signal: it stops a Serve route to another T3 server from counting as this server's endpoint. No unrelated change was found. Mismatch: the PR description says resolveServerAuthPolicy, managedLimits.ts, and its test were changed. The diff against the merge base does not include those files, so those described changes are absent at the reviewed head. This is a description/diff mismatch, not an out-of-scope change.

Full details: Description check

Explanation

The description follows the template's main sections (Problem, Change, Scope, Verification) and gives substantial detail. It names the problem (browser Settings shows 'Only this machine can connect' for a t3 serve --tailscale-serve server), links issue #35521, explains the web-only change in isServerNetworkAccessible, and lists reproduction steps, before/after evidence, and test results. Gaps: the Problem section does not have a clear standalone heading-level problem statement separate from the reproduction notes, and the Scope section does not explicitly say why the change is a bug fix rather than a feature. The description also contains contradictory and stale content. It says the branch adds nothing on the server, yet it lists a server test case and a before/after comparison against commits that describe server-side changes. The 'Not in this change' list repeats the same bullet about /api/auth/session twice. The Verification section cites tests run against the server package, which the description says the branch no longer changes. The title is not provided in the supplied context, so the conventional commit title requirement cannot be checked. The template also asks for the agent model and harness; the AI-assisted note does not name them.

Resolution

Clarify the scope of the change. State whether the branch changes server code. If it does not, remove or relabel the server test and the 'EnvironmentAuthPolicy.test.ts' reproduction steps that describe a server change. Remove the duplicated bullet in 'Not in this change'. Make the Scope section say why this is a bug fix and link the issue. Confirm the PR title uses a conventional commit format. Name the model and harness used, as the template requests.

Finishing Touches

Generate unit tests (beta)

• Commit to this branch
• Create a new PR

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

Recent review info

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: factory/issue-35521
• File: .coderabbit.config.ts
• Commit: d0aaee1
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Team
• Run ID: 252384af-cd30-4bce-87ef-d896541d28b2

Commits

Reviewing files that changed from the base of the PR and between bf801eb and d0aaee1.

Files selected for processing: 4

• apps/server/src/environment/DirectEndpoints.test.ts
• apps/server/src/environment/DirectEndpoints.ts
• apps/server/src/server.ts
• packages/tailscale/src/tailscale.ts

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

[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	21.5 KiB	21.5 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	1	1	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	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	1	1	0 (0.0%)	8	✅

Other scenarios

All 64 scenario measurements match main.

All 64 scenario measurements

Provider	Scenario	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	short turn resume	WebSocket decoded	10.5 KiB	10.5 KiB	0 B (0.0%)	12.7 KiB	✅
Codex	short turn resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	long turn running, cold open	HTTP decoded	281.7 KiB	281.7 KiB	0 B (0.0%)	302.7 KiB	✅
Codex	long turn running, cold open	rows	75	75	0 (0.0%)	75	✅
Codex	long turn running, history page	HTTP decoded	92.4 KiB	92.4 KiB	0 B (0.0%)	107.4 KiB	✅
Codex	long turn running, history page	rows	75	75	0 (0.0%)	75	✅
Codex	long turn finished, cold open	HTTP decoded	381.5 KiB	381.5 KiB	0 B (0.0%)	410.2 KiB	✅
Codex	long turn finished, cold open	rows	151	151	0 (0.0%)	151	✅
Codex	long turn finished, socket cold open	WebSocket decoded	381.6 KiB	381.6 KiB	0 B (0.0%)	410.2 KiB	✅
Codex	long turn finished, socket cold open	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	long turn finished, socket cold open	rows	151	151	0 (0.0%)	151	✅
Codex	long turn finished, work page	HTTP decoded	186.2 KiB	186.2 KiB	0 B (0.0%)	200.2 KiB	✅
Codex	long turn finished, work page	rows	150	150	0 (0.0%)	150	✅
Codex	idle thread resume	WebSocket decoded	65 B	65 B	0 B (0.0%)	200 B	✅
Codex	idle thread resume	WebSocket messages	1	1	0 (0.0%)	1	✅
Codex	fleet shell snapshot	HTTP decoded	41.6 KiB	41.6 KiB	0 B (0.0%)	53.7 KiB	✅
Codex	fleet shell snapshot	rows	27	27	0 (0.0%)	27	✅
Codex	fleet shell subscribe	WebSocket decoded	65 B	65 B	0 B (0.0%)	200 B	✅
Codex	fleet shell subscribe	WebSocket messages	1	1	0 (0.0%)	1	✅
Codex	fleet shell resume check	HTTP decoded	18 B	18 B	0 B (0.0%)	100 B	✅
Codex	fleet shell resume check	rows	0	0	0	0	✅
Codex	stale shell reload	HTTP decoded	41.6 KiB	41.6 KiB	0 B (0.0%)	53.7 KiB	✅
Codex	stale shell reload	rows	27	27	0 (0.0%)	27	✅
Codex	stale shell resume	WebSocket decoded	140 B	140 B	0 B (0.0%)	400 B	✅
Codex	stale shell resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	stale todo list cold open	HTTP decoded	72.4 KiB	72.4 KiB	0 B (0.0%)	83.0 KiB	✅
Codex	stale todo list cold open	rows	75	75	0 (0.0%)	75	✅
Codex	stale todo list cold open	nodes	2	2	0 (0.0%)	4	✅
Codex	stale todo list cold open	subagents	2	2	0 (0.0%)	4	✅
Codex	stale todo list cold open	plans	1	1	0 (0.0%)	2	✅
Codex	stale todo list cold open	providerTurns	1	1	0 (0.0%)	2	✅
Codex	stale todo list cold open	checkpoints	1	1	0 (0.0%)	2	✅
Claude	short turn resume	WebSocket decoded	10.6 KiB	10.6 KiB	0 B (0.0%)	12.7 KiB	✅
Claude	short turn resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	long turn running, cold open	HTTP decoded	283.3 KiB	283.3 KiB	0 B (0.0%)	302.7 KiB	✅
Claude	long turn running, cold open	rows	75	75	0 (0.0%)	75	✅
Claude	long turn running, history page	HTTP decoded	93.4 KiB	93.4 KiB	0 B (0.0%)	107.4 KiB	✅
Claude	long turn running, history page	rows	75	75	0 (0.0%)	75	✅
Claude	long turn finished, cold open	HTTP decoded	384.2 KiB	384.2 KiB	0 B (0.0%)	410.2 KiB	✅
Claude	long turn finished, cold open	rows	151	151	0 (0.0%)	151	✅
Claude	long turn finished, socket cold open	WebSocket decoded	384.3 KiB	384.3 KiB	0 B (0.0%)	410.2 KiB	✅
Claude	long turn finished, socket cold open	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	long turn finished, socket cold open	rows	151	151	0 (0.0%)	151	✅
Claude	long turn finished, work page	HTTP decoded	188.1 KiB	188.1 KiB	0 B (0.0%)	200.2 KiB	✅
Claude	long turn finished, work page	rows	150	150	0 (0.0%)	150	✅
Claude	idle thread resume	WebSocket decoded	66 B	66 B	0 B (0.0%)	200 B	✅
Claude	idle thread resume	WebSocket messages	1	1	0 (0.0%)	1	✅
Claude	fleet shell snapshot	HTTP decoded	42.2 KiB	42.2 KiB	0 B (0.0%)	53.7 KiB	✅
Claude	fleet shell snapshot	rows	27	27	0 (0.0%)	27	✅
Claude	fleet shell subscribe	WebSocket decoded	66 B	66 B	0 B (0.0%)	200 B	✅
Claude	fleet shell subscribe	WebSocket messages	1	1	0 (0.0%)	1	✅
Claude	fleet shell resume check	HTTP decoded	18 B	18 B	0 B (0.0%)	100 B	✅
Claude	fleet shell resume check	rows	0	0	0	0	✅
Claude	stale shell reload	HTTP decoded	42.2 KiB	42.2 KiB	0 B (0.0%)	53.7 KiB	✅
Claude	stale shell reload	rows	27	27	0 (0.0%)	27	✅
Claude	stale shell resume	WebSocket decoded	140 B	140 B	0 B (0.0%)	400 B	✅
Claude	stale shell resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	stale todo list cold open	HTTP decoded	73.4 KiB	73.4 KiB	0 B (0.0%)	83.0 KiB	✅
Claude	stale todo list cold open	rows	75	75	0 (0.0%)	75	✅
Claude	stale todo list cold open	nodes	2	2	0 (0.0%)	4	✅
Claude	stale todo list cold open	subagents	2	2	0 (0.0%)	4	✅
Claude	stale todo list cold open	plans	1	1	0 (0.0%)	2	✅
Claude	stale todo list cold open	providerTurns	1	1	0 (0.0%)	2	✅
Claude	stale todo list cold open	checkpoints	1	1	0 (0.0%)	2	✅

Baseline: a7779fc · PR result: c7051be · 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.

Dismissing prior approval to re-evaluate bf801eb

Dismissing prior approval to re-evaluate d0aaee1

Dismissing prior approval to re-evaluate 6108604