- snapshot-refs.ts: one check for what live things still boot from (goldens of any live status, pool
images, active environment builds, live workspaces' own snapshots, and a live workspace's boot image
until it has a snapshot of its own).
- Retired goldens beyond the rollback and deleted environments' builds are marked deleted only after the
provider delete succeeded (a failed delete used to be marked and leaked for good), and are deferred
while in use. A new hourly housekeeping step (sweepReleasedSnapshots) retries both.
- Team delete marks the team's environments deleted, so their build snapshots are freed (they leaked).
- Environment prune skips a build a live workspace still boots from.
- Orphan sweeps (per box and project-wide automatic) also respect pool and boot-image references.
- The golden smoke box is non-persistent: it no longer leaves a ~325 MB automatic snapshot behind.
Belt and braces for the capability gate: an auth that advertises console.billing_hint but refuses the fields (rollback, early flag) still mints a ticket, so the console opens signed in without the preselect.
"Continue in browser" sends the plan and billing period picked in the plans dialog with the
console ticket ({page:'billing', plan, interval}) so the web billing page can preselect them. Main
adds the fields only when auth advertises the `console.billing_hint` capability: the ticket route
takes a strict body, and an older auth would refuse them and lose the signed-in landing. The
renderer's hint is re-validated in main (plan id pattern, month|year, billing page only).
The build sheet's no-plan blocker, after Compare plans, no longer says "Finish in your browser":
it says to choose a plan and check again, with See plans to reopen the dialog. Update payment
method keeps going straight to the console.
"Plan & billing", "Upgrade to Orca Cloud", "Choose plan" and every "Compare plans" now open an
in-app plans dialog first: the team's plan and status (trial hours left, active with hours and
seats, payment failed, or no plan with the trial billing offers), then billing's plan cards (Team
with a monthly/annual toggle, Enterprise with contact sales). "Continue in browser" opens the
console's billing page through the existing sign-in ticket. Members see the same plan info with
"Ask a team admin to upgrade" (Plan & billing is back in their account menu, read-only). Read-only:
no checkout, usage history or admin settings return to the desktop.
Everything shown is billing's: names, Stripe prices, seat pricing, feature bullets and trial terms
from GET /v1/plans, now also each plan's tagline, feature-list lead and button label (new optional
catalog `marketing`, unversioned copy, plus `omitFeatures` so Enterprise stops advertising egress
controls). Main's lenient parse keeps a newer or older service safe; without the catalog the
entitlement's offers stand in, and with neither the dialog says plans didn't load and still offers
the console. While billing loads the title is neutral, so a paying team never sees an upsell flash.
Settings > Account's billing row opens the dialog ("View plans"); "Update payment method" still
goes straight to the console.
The desktop's Routines settings are behind CLOUD_ROUTINES_UI_ENABLED=false, so the CLI
(orca cloud routine run) is where people read a run's result. A run that failed with Claude's
own "Not logged in · Please run /login" now says Claude isn't set up for Orca Cloud and where to
set it up, and doesn't repeat the raw summary. The detection moves to src/shared so both surfaces
use one rule.
- Plan & billing (Settings › Account and the account menu, including Upgrade) shows only to an
owner or admin of the active team: the console's own rule (canManageMembers on
/console/orgs/:id/billing). Members were offered a row the console then refused.
- An archived workspace's terminal bar says it's archived and offers Unarchive (was: asleep, Wake).
- A routine run whose agent exited with Claude's "Not logged in · Please run /login" says Claude
isn't set up for Orca Cloud and where to set it up, instead of the CLI's text.
- A Sleep or Archive the person chose holds the terminal reconnect notices from the press until a
wake starts (or 90 s): a paired workspace's status lags the sleep call, and in that gap a pane
said "Reattaching terminal · Down for 10 sec".
On Orca Cloud for a GitHub-only project the composer passed its name field as
the create's prompt: it titled the workspace (NAME-DROPPED, #563) but was also
the quickPrompt and the first agent's prompt, and with Blank terminal (agent
null) the default agent was resolved and sent the name as its first message.
- StartFromGitHubRequest / GitHubCreateSource / useCloudOnlyCreate: prompt -> name
- githubCreateRequest: title from the name as before, quickPrompt is empty
- launchFirstCloudAgent: no prompt; a chosen TUI agent starts with none
(PAIRED-HELD-STRIPS, #568), Blank terminal starts nothing
- Back to draft keeps restoring the typed name
Co-Authored-By: Claude <noreply@anthropic.com>
One Checkout sends checkout.session.completed, customer.subscription.created and invoice.paid,
often concurrently; each that re-read the row before another wrote it saw the plan start and
pushed billing.plan_started with its own 'at', so the event ids differed and the team's audit
log showed it twice. A plan starting or ending now carries a once key (subscription id and
period start), so every racer pushes the same event id and the control plane's INSERT OR IGNORE
keeps one.
After #738 both pages were a single "Manage in browser" card. Team and admin now have no desktop
page: Settings › Account shows one row, "Team & admin", that opens the console home through the
single-use ticket, beside the existing Plan & billing row. The account menu keeps Team and
Plan & billing (both open the console). Searching for members, SSO, the audit log or Enterprise
lands on the Account row. The 'cloud-team' and 'cloud-enterprise' settings targets, their search
entries and their strings are removed; nothing else linked to them.
Owner decision: paid and admin settings do not belong in the open-source desktop app.
Deletes the Enterprise pane (SSO, domains, SCIM, MFA policy, audit log), the Plan & billing
section and plans dialog (checkout, usage history, plan-feature-copy), and Team management
(members, invites, roles, leave, delete, create), with their hooks, fixtures, tests, i18n keys
and the main/preload/IPC only they used.
Each removed section is now one "Manage in browser" row. Main asks auth for a single-use
console code (POST /v1/desktop/console/ticket {page, orgId?} -> {code}) and opens
<consoleOrigin>/console/<section>?code=... with shell.openExternal; on any failure it opens the
page without a code. Endpoint, field names and origin live in cloud-console-config.ts
(ORCA_CLOUD_CONSOLE_URL overrides the origin, default is the auth origin).
Kept: sign-in, account identity and own MFA row, the team MFA gate/banners, the team switcher,
the roster (presence, names), plan status badge and payment notice, in-context billing
blockers (now "Compare plans" opens the console and re-checks on focus), cloud setup.
POST /v1/billing/checkout and /portal accept returnTo: 'console'. The return URL is built from
this Worker's CONSOLE_URL (else the setup pages' auth origin, which forwards /console) and the
caller's own team, never from the request. Desktop-started flows keep the /return handoff page.
The SSO line said "SSO, SCIM provisioning and audit logs" next to billing's own "Team audit log" line; it now reads "SSO and SCIM provisioning". Copy only, no entitlement change.
The audit log scroller uses the shared useScrollEdgeFades mask: a top fade only once scrolled, a bottom fade only while more is below. The rule and focus ring move to an unmasked frame.
* fix(settings): audit log scrolls in place instead of growing the settings page
The rows and Show more sit in a keyboard-focusable region capped at min(24rem, 60vh); the header, export and filters stay above it.
* style: oxfmt
* fix(settings): add the audit scroll label to en.json; move the scroll region to its own file (max-lines)
* fix(plans): stop advertising egress controls on the Enterprise plan
No team can choose an egress policy today (SANDBOX_EGRESS_PRESET is operator-wide and open by default), so the plan cards drop the egress_controls line. The catalog feature and entitlements are unchanged.
* style: oxfmt
The root pnpm typecheck covers no cloud/ package and vitest strips types, so cloud
packages' own type errors went unseen (261 in sandbox-control). premerge-check.sh now
runs the own pnpm typecheck of every touched cloud/apps or cloud/packages package listed
in premerge-typecheck-packages.txt, after installing the cloud workspace and building its
packages. The list holds the 11 packages that are clean today; sandbox-control joins once
its errors are fixed.
* fix(cloud): Orca Cloud on a remote-server project whose GitHub repository it can't reach says so with More access, not the folder wording (UNREACHABLE-REPO-NO-ROUTE)
* fix(cloud): UNREACHABLE-REPO-NO-ROUTE design-squad: 'Give Orca access' link, muted (not red) precondition line, fixture
* test(cloud): OrcaCloudNeedsRepoSubmit fixtures press the footer's Create
* fix(cloud): UNREACHABLE-REPO-NO-ROUTE the Project field's ring is amber while Orca Cloud waits for a GitHub repository (design-squad); aria-invalid kept
* fix(cloud): UNREACHABLE-REPO-NO-ROUTE ring classes stay literal (localization audit), drop the unreferenced remoteServerDetail key
* fix(cloud): a refused create's row offers Retry and Dismiss, and a new create replaces it (REFUSED-CREATE-ROW-STICKY)
* fix(cloud): REFUSED-CREATE-ROW-STICKY a billing or limit refusal's cloud glyph is amber, a real failure stays red (design-squad)
* fix(cloud): a project cloned only on a remote server offers Orca Cloud in Run on and starts it from its GitHub repository (RUNON-NO-NEW-SANDBOX)
* fix(cloud): RUNON-NO-NEW-SANDBOX design-squad wording for an unreachable repository
The team-switch leave fired only for an owned workspace's host going from
the team's (+) to hidden (-); a paired one's host is ~ since #690, so the
main pane kept showing another team's workspace after the switch (n57).
It leaves for ~ to - too. The status bar's host list and count now drop
another team's cloud servers, the same scope as the sidebar (#722).
useCloudRecoveryContext counted only owned workspaces as cloud, so for a
paired one its asleep and waking states were unknown to the recovery
notices: the remote-server outage notice ('Reconnecting to ... (cloud)',
role=status, aria-live=polite) stayed up under the asleep view, where a
screen reader would announce it (n56 DOM dump). Paired rows count as cloud
there too, so the notice stays silent while asleep or waking.
Right after pairing, the server can answer the project and worktree reads
before it lists the workspace's own row; the row lands a moment later.
connectAndReveal looked once, found nothing, toasted 'is connected' and
returned, so the picker closed and nothing was selected (n48, n55). It now
waits a bounded 15 s for the row (the picker spinner and #710's Opening
toast cover the wait), then reveals it as before.