Quote SCP operands when building commands to prevent shell injection through source or destination paths, and cover the escaping behavior in the SSH command injection tests.
Drop the unused GitHub release and version sync options from sync:bunny,
leaving the command focused on BunnyCDN template, release, and nightly syncs.
Update the nightly test to assert it does not invoke gh or git.
Check the selected S3 storage against the database at submit time so
stale Livewire state cannot schedule backups with storage that was
reassigned or marked unusable after the component mounted.
Prevent scheduled database backups from enabling S3 uploads without a valid team-owned storage configuration, and preserve the previous S3 storage ID in missing-storage error messages.
Add coverage for backup edit/create validation and S3 upload failure messaging.
Wrap first-use SSH and SCP multiplexed commands with a lock to avoid racing while the control socket is created. Also detect native OpenSSH mux master process names during stale connection cleanup and cover both orphaned and duplicate mux processes with tests.
Remove explicit SSH master pre-warming and lock handling so OpenSSH manages ControlMaster creation lazily from real ssh/scp commands. Add cleanup for duplicate mux processes and update coverage around mux command options and stale process cleanup.
Serialize multiplexed SSH master creation per server to avoid concurrent workers spawning orphaned processes. Enable scheduled cleanup for stale mux connections and add guarded orphan process reaping with tests.
- resolve the GitHub App by a stable identifier during installation
callbacks so installing and re-installing keeps working over the
full lifetime of the App
- verify the installation id received from the callback against the
GitHub API before persisting it
- support re-installing an already configured GitHub App instead of
blocking it
- require an authenticated session and rate limit the setup callback
routes
- extend manifest setup state validity to match GitHub's manifest
code lifetime
Adds feature coverage for the GitHub App setup and installation
callbacks.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Centralize scheduled job queue selection with crons_queue() and use it for scheduler, task, and database backup jobs so cloud runs on crons while self-hosted stays on high.
Use config-based queue selection for deployment and scheduled jobs so cloud dispatches deployments to `deployments` and scheduled jobs to `crons`, while self-hosted keeps using `high`.
Add coverage for deployment queue helper, start action routing, and scheduled job manager routing.
Git hosts treat owner/repo names case-insensitively, but the exact
repository match used a case-sensitive comparison, so a payload whose
casing differed from the stored git remote would fail to match and
skip a legitimate deployment.
Lowercase both canonical repository paths before comparing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The manual webhook handlers selected target applications with a
`git_repository LIKE %full_name%` substring query, so a payload
repository name could match unintended applications when repository
names overlap.
Add a `MatchesManualWebhookApplications` trait that validates the
incoming `owner/repo` value and matches `Application.git_repository`
by exact normalized path. Github, Gitlab, Gitea and Bitbucket manual
handlers now use it, reject invalid repository input early, and return
a consistent generic webhook failure payload.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Resolve the server and network in Destination::addServer() and
::promote() through ownedByCurrentTeam() before use, authorize the
update against the resource, and pass the validated IDs into
attach()/detach()/update(). Errors are routed through handleError()
to match the sibling removeServer() method.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Move Sentinel push handling into a controller and dispatch server update jobs only when container state changes or the force interval elapses. Add opt-in PostgreSQL read/write replica configuration and tune periodic proxy network and storage checks to reduce unnecessary work.
Add feature coverage for replica config, Sentinel push deduplication, deployment log scrolling, and server update job optimizations.
Scope DeleteEnvironment::mount() and delete() lookups through
Environment::ownedByCurrentTeam() so an environment_id that belongs to
another team resolves to a 404 instead of loading the foreign record.
Mark $environment_id as #[Locked] so the public Livewire property can no
longer be reassigned from the client.
Add tests/Feature/DeleteEnvironmentTeamScopingTest.php covering mount,
delete, the #[Locked] guard, and the team-scoped helper for both the
cross-team and own-team cases.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The Source component now resolves the supplied private key and Git
source IDs through team-scoped queries before persisting them, so a
selection can only ever reference a resource owned by the current
team. The source type is additionally restricted to the supported
GitHub/GitLab app classes.
The privateKeyId property is marked #[Locked] so it can only change
through the dedicated handler rather than a direct property update.
Adds feature tests covering team-scoped selection of private keys and
Git sources.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Clickhouse, Dragonfly, and Keydb still called syncData() inside the
DatabaseProxyStopped broadcast handler, clobbering in-progress edits to
name/description/credentials. Refresh only is_public/public_port/
public_port_timeout instead, matching the pattern used elsewhere.
Also null-guard HasDatabaseStatusInfo::getListeners() against an absent
Auth::user()/currentTeam(), add explicit return types on getListeners()
and render(), and convert inline comments in the SSL refresh test to a
PHPDoc block.
The earlier refreshStatus fix kept user-typed values intact but Livewire still
absorbed deferred wire:model values into the snapshot on every broadcast-
triggered roundtrip, clearing the unsaved-changes indicator and making the form
look auto-saved. Move all status-derived display (DB URLs, SSL toggle/mode,
cert expiry) out of each DB General form into a sibling StatusInfo Livewire
component, so the form never roundtrips on broadcasts.
Shared scaffolding lives in App\Traits\HasDatabaseStatusInfo plus an x-database-
status-info Blade component, leaving each per-DB StatusInfo class as a ~20-50
line declaration of label, SSL mode options, and SSL save hooks. Parents
dispatch databaseUpdated from save methods so the sibling refreshes after writes.
Tests cover the architecture (no DB form subscribes to status broadcasts) and
the sibling's refresh-on-status-change behavior.
Build the global destinations list from actual destination records so empty
servers do not render duplicate empty states. Allow creating Docker destinations
for a selected team server outside the global usable list, authorize swarm
creation correctly, and store discovered swarm network names from the selected
network. Add feature coverage for empty states, selected-server mounting, and
swarm destination creation.
Exclude generated Docker Compose SERVICE_FQDN, SERVICE_URL, and SERVICE_NAME variables from runtime, build-time, and build arg environments so stale stored values cannot override generated service names for preview deployments.
Call the preview environment variable relationship as a query when building the legacy configuration hash, and cover preview deployments with a regression test.
Dispatch configuration change events after saving application source and advanced settings, and refresh the configuration checker before showing redeploy diffs.
Store deployment configuration snapshots on application deployment queues and compare them against the current application state. Surface grouped pending changes in the configuration checker and use build-impact diffs to decide when an existing image can skip the build step.
Add centralized stop grace period resolution for application settings and use it across manual stops, preview stops, and deployments. Validate the Livewire advanced setting against shared min/max constants and cover persistence, fillable creation, and fallback behavior with tests.
Decode base64 custom_nginx_configuration before model assignment so it is not double-encoded, and allow null values when clearing the setting. Add API coverage for create, update, invalid input, and clearing behavior.
Generate pull request preview image tags with both the PR id and commit
so different commits on the same PR do not reuse the same image tag. Sanitize
and truncate generated tags to stay within Docker tag limits.
Require Docker buildx before Railpack builds, normalize environment
variable keys before validation, and align private deploy key API docs with
the supported dockerfile build pack.
Avoid failing Stripe webhook processing when local subscriptions are missing, and cover ignored invoice/payment/subscription events with feature tests.
Add shared environment variable key validation and normalization for Livewire forms and models, allowing Docker-compatible keys while rejecting invalid entries such as keys containing equals signs. Also quote Railpack build environment and secret arguments safely.
Build Railpack variables from generic build-time vars plus Railpack-specific vars, filter unrelated buildpack control vars, and ensure curl/wget deploy apt packages are present. Add coverage for standard and preview deployments.
Drop the unstable applications/dockercompose route and controller path now that
service creation is handled by POST /api/v1/services. Add coverage to ensure the
deprecated endpoint stays unregistered while the services endpoint remains
available.
- `/mcp/enable` and `/mcp/disable` now use POST (state-mutating ops)
- `ListServices` queries DB directly instead of loading all projects into memory
- `ListApplications` validates tag arg rejects empty string (not just falsy)
Move copyLogs from PHP Livewire method to Alpine.js to avoid
unnecessary server round-trips. Extract collectVisibleLogs()
helper shared by both copy and download actions.
Hide refund section entirely when not eligible instead of
rendering a permanently disabled button.
Replace nested null checks with nullsafe operator chains, add ?Server
return type, drop unreachable database branch, and cover all paths with
feature tests.
Mirrors Nixpacks behavior: inject COOLIFY_* and SOURCE_COMMIT into
railpack build variables so apps (e.g. SPAs baking public URLs) can
read them via /run/secrets/<KEY>.
Rename is_nixpacks → is_buildpack_control to cover both NIXPACKS_ and
RAILPACK_ prefixed keys. Update the env variable view and appends list
accordingly.
Promote generate_coolify_env_variables to protected for testability.
Replace `railpack_environment_variables_collection()` helper (which returned
pre-filtered Eloquent attribute collections) with inline queries on
`environment_variables()` / `environment_variables_preview()` filtered by
`is_buildtime`. This ensures Railpack build variables are sourced from the
same query path as the rest of the deployment pipeline and avoids relying on
a now-removed accessor that silently included all railpack vars regardless of
build context.
Remove RAILPACK_FRONTEND_IMAGE env var from helper Dockerfile and resolve
the image ref at runtime using a new `railpack_version` constant in config.
Eliminates Docker build-time env interpolation for BUILDKIT_SYNTAX arg.
Replace random string with encrypted JSON payload containing
server_uuid, binding token to its server for validation.
Remove double-encrypt test no longer relevant to new token format.
Replace hard validation error with self-healing token logic. Tokens that
are null, empty, or fail decryption are now regenerated automatically
rather than crashing sentinel startup or metrics reads.
Token format changed from encrypted JSON payload to a plain 64-char
random string (Str::random), eliminating double-encryption issues and
simplifying the validation regex to cover the new character set.
New `ensureValidSentinelToken()` method on ServerSetting centralises
the get-or-regenerate contract; both StartSentinel and HasMetrics now
delegate to it. HasMetrics logs a warning when regeneration occurs so
operators know a sentinel container restart is required.
`isValidSentinelToken()` now accepts `?string` (null → false).
Adds feature tests covering: null/empty/undecryptable stored values,
idempotent return of valid tokens, RuntimeException only when
regeneration itself produces an invalid token, no double-encryption of
newly generated tokens, and cast round-trip consistency.
On cloud, DeploymentException and TimeoutExceededException are expected
failure modes that pollute the Horizon failed jobs UI. Listen to JobFailed
events and scrub the entry via JobRepository::deleteFailed so operators
are not alerted for noise failures. Self-hosted instances are unaffected.
Add Model Context Protocol server exposing Coolify infrastructure data
to AI assistants. Includes tools for listing/fetching servers, projects,
applications, databases, and services, scoped to authenticated team tokens.
- Add CoolifyServer with 10 read-only tools (list/get for all resource types)
- Add BuildsResponse and ResolvesTeam traits for shared tool logic
- Add EnsureMcpEnabled middleware guarding /mcp routes
- Add enable/disable MCP API endpoints (root-only)
- Add is_mcp_server_enabled toggle in instance settings and advanced UI
- Add migration for is_mcp_server_enabled column
- Add feature tests for MCP endpoints and toggle API
- Scrub sensitive keys (passwords, tokens, raw IDs) from all responses
Add DetectsSkipDeployCommits trait with two strategies: shouldSkipDeploy
(all commits must contain the marker) for push events, and
shouldSkipDeployAny (any single marker triggers skip) for PR/MR titles
and latest-commit signals.
Apply trait to Bitbucket, Gitea, GitHub, GitLab webhook controllers and
ProcessGithubPullRequestWebhook job. PRs pass pullRequestTitle through
to the job constructor for evaluation.
Large host files mounted via Docker volumes caused the storages page to
become unusable — full file content was stored in the encrypted mediumText
column and serialised into the Livewire payload, crashing the browser.
- Add MAX_CONTENT_SIZE (5 MiB), BINARY_PLACEHOLDER, and TOO_LARGE_PLACEHOLDER
constants to LocalFileVolume
- Check remote file size via stat/wc before cat in loadStorageOnServer and
saveStorageOnServer; store placeholder instead of content when limit exceeded
- Expose is_too_large computed attribute (appended for Livewire serialisation)
- Guard submit, instantSave, and syncData in FileStorage Livewire component
- Truncate oversized content in Storage::refreshStorages to prevent payload bloat
- Show distinct warning banner in file-storage blade; mark textarea readonly and
hide Save/Convert buttons for too-large files
- Add unit tests covering constants, computed flags, and toArray serialisation
Fixes#4701
Add `connection_timeout` field to server settings, allowing per-server
override of the global SSH connection timeout constant.
- Migration adds `connection_timeout` integer column (default 10s)
- `ServerSetting` model exposes and casts the new field
- `SshMultiplexingHelper::getConnectionTimeout()` resolves per-server
value with fallback to `constants.ssh.connection_timeout`
- All SSH/SCP command builders use the new resolver instead of the
global config directly
- Livewire `Show` component binds `connectionTimeout` with validation
(1–300 seconds) and syncs to/from the model
- UI input added to server settings form with helper text
- Feature tests cover default, persistence, resolver, and fallback
Move reachability notification triggering out of isReachableChanged into
a dedicated ServerReachabilityChanged event dispatched by
ServerConnectionCheckJob. Remove the blocking 3-attempt sleep loop from
isReachableChanged — unreachable_count threshold alone now gates the
Unreachable notification. Add feature and unit tests covering all
notification dispatch paths.
Introduce a dedicated `audit` log channel (daily rotation, configurable retention via
LOG_AUDIT_DAYS) and a small `auditLog()` / `auditLogWebhookFailure()` helper used to
record state-changing API operations and webhook events.
Instrumented:
- API mutation endpoints (create / update / delete / start / stop / restart) across
applications, services, databases (incl. backups, env vars, storage), servers,
projects + environments, scheduled tasks, private keys, GitHub apps, cloud provider
tokens, Hetzner server provisioning, instance enable/disable.
- Webhook signature verification outcomes for GitHub, GitLab, Bitbucket, Gitea and
Stripe, plus the Sentinel push endpoint.
- Authentication and authorization outcomes via the global exception handler and
the `ApiAbility` middleware (unauthenticated, ability-denied, policy-denied).
The helper is wrapped in try/catch so logging failures never affect the request
path. Successful operations log at `info`; suspicious/denied requests log at
`warning`. Operators wanting a failures-only feed can set `LOG_AUDIT_LEVEL=warning`.
Includes a feature test suite covering the helper, the webhook providers and the
new auth/authorization log paths.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Extract NIXPACKS_/RAILPACK_ prefix filtering into a reusable
`scopeWithoutBuildpackControlVariables` query scope on EnvironmentVariable.
Apply scope consistently to runtime vars, runtime preview vars, and
buildtime var generation in ApplicationDeploymentJob.
Refactor `generate_railpack_env_variables` to return a Collection.
Add `RAILPACK_FRONTEND_IMAGE` constant and bake it into the
coolify-helper Dockerfile as a build arg.
Add DevelopmentRailpackExamplesSeeder (dev/local env only) for
seeding example Railpack apps, wired into DatabaseSeeder.
Add tests:
- ApplicationDeploymentControlVarFilteringTest: verifies control vars
are excluded from runtime and buildtime envs
- DevelopmentRailpackExamplesSeederTest: verifies seeder behavior
- ApplicationDeploymentRailpackEnvParityTest: parity checks for env
handling across build/runtime paths
Dockerfile buildpack was passing --network {custom_network_name} to
docker build, but BuildKit only supports host, none, and default.
Every other buildpack already uses --network host with --add-host
flags. Aligned the Dockerfile path to match.
Fixes#9804
Adds stop_grace_period to application settings (seconds, 1-3600, default 30).
Used in place of the hardcoded docker stop -t 30 in the four places that
stop application containers: rolling update shutdown, manual stop, stop on
another server, and preview deployment stop.
Non-positive values fall back to the default via ($val > 0) ? $val : default,
with tests covering 0 and -10 so the cast does not blow up if a bad value
ever lands in the db.
Picks up Jack Coy's work from #7125 which went dormant. His commits are
squashed here with credit below.
Co-authored-by: Jack Coy <jackman3000@gmail.com>
Replace the flat character-class regex for SHELL_SAFE_COMMAND_PATTERN with
a token-aware alternation. The parser now recognizes explicit tokens
(`&&`, `||`, balanced single/double quotes, whitespace, and an unquoted
safe-char run) instead of a bag of characters, which lets us extend the
accepted grammar without loosening the guarantees.
New surface area, with tests:
- logical OR chaining (`make build || make clean`)
- shell globs and bang (`rm *.tmp`, `cp src/?.js dist/`, `! grep -q foo`)
- single-quoted arguments are now treated as balanced runs rather than
rejected per-character
Preserved surface area:
- && chaining, balanced "..." and '...' quotes, the previous safe path /
argument characters, and the existing error-path contract in
ApplicationDeploymentJob::validateShellSafeCommand().
Also refreshes the user-facing validation messages in General.php so the
allow/deny list shown on failure matches the new grammar.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
