fix(restore): drop USE/CREATE DATABASE statements so mysql/mariadb restores target the selected database
(cherry picked from commit ce0fe63b4a)
[skip ci]
Organization admins can define their own AI providers (name + API URL)
from the AI settings section. When at least one custom provider is
defined, it replaces the built-in provider list in the Add AI form,
the API URL is auto-filled and locked, and the backend rejects any
configuration whose URL is not in the allowed list.
The verify-totp handler checked for the error code
INVALID_TWO_FACTOR_AUTHENTICATION, which no longer exists in
better-auth 1.6.23 (the two-factor plugin now returns INVALID_CODE).
As a result, entering a wrong TOTP code fell through to the generic
catch and showed "Error verifying 2FA code / Unknown error" instead of
a clear "Invalid verification code" message.
Match the current better-auth error code so the specific, actionable
message is shown.
The organization switcher's DropdownMenuContent inherited its width from
the collapsed trigger (~40px) via the base primitive's
w-(--radix-dropdown-menu-trigger-width), clamping the menu to the
min-w-32 (128px) floor. This cut off the org name and action buttons in
icon mode.
Set an explicit w-64 so the menu fits its content regardless of the
trigger width, matching the pattern already used by the notification
dropdown in the same file.
Fixes#4840
The remote (execAsyncRemote) path built `echo ${password} | docker ${args.join(" ")}`
with the password, registryUrl and username interpolated unescaped, so a password
like `pw; whoami` ran arbitrary commands as root on the target server. Reuse
safeDockerLoginCommand (already used by create/update), which shell-escapes each
field and feeds the password via --password-stdin. The local argv+stdin path was
already safe.
server.remove deleted a server (and its deployment rows) by caller-supplied
serverId without checking it belongs to the active organization, unlike server.one
and server.update. An owner/admin of org A could delete org B's server registration.
Resolve and compare the server's organizationId before the active-services guard,
so cross-org callers are rejected without leaking existence.
Add VOLUME_NAME_REGEX (Docker volume-name format) and enforce it on volumeName in
create/update/runManually — a legit volume name never contains shell metacharacters,
so this blocks injection across every docker run/rm/rclone sink at once. Escape the
user-supplied backupFileName and the mount volumeName in database rebuild with quote().
Wrap S3 credential flags (getS3Credentials + destination.testConnection), the
listBackupFiles search path, and every restore backupPath/backupFile with
shell-quote's quote() so $(), backticks, quotes and spaces in destination
fields or backupFile can no longer break out of the rclone shell commands.
Use shell-quote's quote([...]) directly at each git-provider clone call site
instead of the one-line shellWord wrapper, and remove the now-unused helper.
Behavior is identical (quote([String(v ?? '')])).
The /api/providers/github/setup callback performed privileged writes (persisting
App client_secret/webhook_secret/private_key, re-pointing installations) with no
session check, deriving the write target (organizationId/userId on gh_init,
githubId on gh_setup) from the attacker-controlled state parameter. An
unauthenticated request, or a member with no gitProviders permission, could plant
or re-point a provider in any organization.
Require a valid session, gate on the gitProviders permission (same guard the tRPC
github router uses), derive the create target from the session, and verify the
gh_setup provider belongs to the caller's organization.
Host-level schedules (server / dokploy-server) run their script as root on the
host. The owner/admin gate only ran in the no-service branch, so a member could
attach an accessible applicationId to a dokploy-server schedule and skip it,
gaining root via schedule.runManually.
Extract assertHostScheduleAccess into the schedule service and call it before
the service-access branch in create/update/delete/runManually so the host-level
authorization always applies.
- application general 'Open Terminal' now passes serviceId (applicationId).
- The docker/terminal DockerTerminalModal now forwards serviceId to its Terminal.
- compose container rows thread serviceId to their terminal + logs modals.
This lets a member with service access (but without the docker permission) open
the terminal / read logs of their own service's containers from every service
view, matching the backend service-first authorization.
When a container is opened from a service page (serviceId present), gate on
service access alone — matching application.readLogs — instead of requiring the
docker permission first. This unblocks a member who has the service but not the
canAccessToDocker permission (or the server it runs on) from reading its logs /
opening its terminal. Generic Docker-overview ops (no serviceId) still require
docker permission + server access.
- Reorder canAccessDockerOverWss: docker permission -> service access (if
serviceId) -> server access, so the service grant is the primary signal.
- Pass serviceId from the service log views (application + postgres/mysql/mariadb/
mongo/redis/libsql + compose) so container logs opened from a service page are
gated by checkServiceAccess, matching the terminal path.
Narciso E. Núñez Arias