* perf: Add cached Integration.findEnabledByTeamIdAndService lookup
Caches a lightweight projection of a team's integrations for a service in
Redis so event processors can cheaply check whether a team uses a service,
e.g. from shouldQueue, before creating a job. The cache is invalidated after
commit by model lifecycle hooks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* chore: Remove comment from Integration cache invalidation
* perf: Skip queueing Slack jobs for teams without a matching post integration
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* fix: Fail fast on Redis for cache and rate limits, lock migrations in Postgres
* fix: Use the fail-fast Redis client for locks
* fix: Require a database pool of at least two connections
* chore: Rename cache Redis client to ephemeralClient
* chore: Move plugin-specific integration code into plugins
Core no longer imports from or names the settings of individual plugins.
Plugins now register mention providers, issue status icons, instance URLs,
settings keys, settings and import schemas, and OAuth clients through hooks.
* fix: Recompute embed instance URLs after plugins load and match legacy issue icons by URL
* fix: Skip logo upload jobs queued without a settings path
* chore: Choose issue status icons from the unfurl state
* perf: Render share page content without the export lock and cache it
* perf: Render share page content only for bots, detect bots with isbot
Browsers never read the server-rendered share content, so skip the render
for them. Replace the koa-useragent isBot check with the isbot package,
which also detects AI crawlers, and add a lint rule against isBot.
* Fix stale workspace entries in sessions cookie
* Use authenticated token expiry for workspace session entries
* Remember logged-out workspaces for three months after last use
* Remove excessive tests
* Fix auth info test to send a real sessions cookie
* chore: Replace tunnel-agent with https-proxy-agent
* fix: Keep the target host in proxied HTTP request URLs
* test: Move test proxy into test support
* chore: Vendorize throng as ClusterManager
* fix: Force kill workers after grace period, skip fork if shutting down
* fix: Give workers their full shutdown timeout before the primary exits
* chore: Use a named export for ShutdownHelper
* chore: Prune redundant comments and normalize lint directives
Removes comments that restate the following line and stale TODOs, drops
the passive event listener feature check now that all supported browsers
support it, and renames eslint directives to oxlint.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BehTjuZwKbN5csMCp3H4QH
* fix: Evaluate restricted team domains at validation time
The restricted email provider list was captured when the model loaded, so
it could not follow the hosting environment and the tests that covered it
never reached the model.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BehTjuZwKbN5csMCp3H4QH
* chore: Replace type suppressions with accurate types
Adds a Concrete helper type for registries of abstract classes, declares
missing Sequelize and refractor typings, uses public ProseMirror APIs, a
typed mock context helper in middleware tests, and a StaticLink component
for sidebar rows being edited.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BehTjuZwKbN5csMCp3H4QH
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: Personal collections backend
Adds an owner to collections. A personal collection is private, one per user
per team, created lazily through collections.personal, and only its owner can
see or manage it. Admin bypasses in policies, lists, exports, events, webhooks
and file operations exclude the personal collections of other users. Turning
off the PersonalDocs preference makes the collection read-only.
* test: Access matrix for personal collections
Checks that collection and document endpoints, list and search results, MCP
tools, collaboration and websocket rooms agree with the read policy for the
owner, collaborators, members, admins, guests and other teams.
* feat: Personal collection in the sidebar
Adds a Personal sidebar section that renders the personal collection with the
collection document tree, creates the collection on the first new document,
and keeps it out of the Collections section. Adds the workspace setting, the
personal sidebar context, and hides collection sharing and membership UI for
personal collections.
* fix: Drag and drop in shared with me trees
Anchors the reorder drop cursor inside the row, and prefetches documents on
hover so that nested rows that were never opened can be dragged.
* feat: Personal collection owner for collaborators, no collection page
Presents the owner of the personal collection on documents, so that
collaborators see the owner in the share popover. Root rows in the Personal
section show the default document icon. A personal collection has no page:
the scene shows not found, and the breadcrumb, command bar and delete redirect
no longer link to it.
* fix: Personal sidebar indent and breadcrumb items without a link
Children of rows at depth 0 skip a level so that they start beyond the row
icon. Breadcrumb items without a destination, such as the personal and deleted
collection, are shown as plain text.
* test: Reduce personal collection tests to list paths
Endpoints that authorize a single model are already covered by the policy
tests. Keeps the lists and searches, which filter in queries, and the
presenter.
* test: Move personal collection list tests into collections tests
* feat: Create documents in the personal collection by alias
documents.create, documents.import and the MCP create_document tool accept
the collectionId "personal", which refers to the personal collection of the
user and creates it when it does not exist yet.
* refactor: Remove collections.personal route
The sidebar creates its first personal document with the personal
collectionId alias, so documents.create is the only path that creates the
personal collection.
* feat: Personal collection name, starring, templates and default
Personal collections cannot be starred, renamed or hold templates, and the
PersonalDocs preference is on by default. The app shows collections through a
displayName getter, which is the translated "Personal" for a personal
collection.
* fix: Use the same policy for starring and unstarring collections
* fix: No pins or viewer insights for personal documents
* revert: Show personal collection events in the audit log
Admins with audit access should see events in personal collections, even when
they cannot open the documents.
* fix: Load collection owner in share scopes
* fix: Keep personal collections out of workspace exports and webhooks
Workspace exports skip all personal collections, while an explicit export of
a personal collection stays available to its owner. Events in personal
collections, including attachments and documents moved after queueing, are
never delivered to workspace webhooks. Personal collections cannot be
renamed or made public through any path, cannot hold pins, and are not
offered as mentions.
* Change icon to avatar
* fix: Personal documents cannot be unpublished
* feat: Personal collection context for MCP
Collections are presented with an isPersonal flag, the server instructions
explain the personal collection, and move_document and update_document accept
the personal collectionId alias.
* fix: Keep moved personal exports private and remove pins on move
Document exports record the owner of the personal collection they came from,
so they stay private after the document moves. Moving documents into a
personal collection removes their pins, including pins to home.
* fix: Only create the personal collection in MCP when it is used
* feat: Support OAuth Client ID Metadata Documents for MCP clients
MCP clients can send an https URL as their client_id. The server fetches
the metadata document at that URL and caches one team-less client row per
URL, so MCP works on the root domain without dynamic registration.
Closes#14010
* fix: Address review findings for client ID metadata documents
- Only count unexpired tokens when cleaning up metadata document clients
- Reject backslashes, whitespace and trailing host dots in client_id URLs
- Strip trailing dots before installation host comparisons
- Type clientSecret as nullable
* fix: Keep OAuth clients with pending authorization codes during cleanup
* refactor: Simplify client ID metadata document handling
- Gate metadata document fetches on the MCP preference inside the helper
- Add isMCPClient getter and a shared client ID prefix constant
- Enforce the team/metadata document CHECK in both directions
- Index oauthClientId on OAuth token and code tables
- Reuse parseDomain, clamp and time helpers
* revert: Remove pending authorization code check from OAuth client cleanup
* fix: Block all special-use addresses when fetching client metadata documents
ALLOWED_PRIVATE_IP_ADDRESSES no longer applies to client ID metadata
document requests (draft -02 section 8.6). The fetch wrapper takes an
allowIPAddressList option to override the environment list.
* fix: Reject empty userinfo in client metadata document URLs
https://@host and https://:@host have a userinfo component that the URL
parser drops, so check the raw authority (draft -02 section 3).
* fix: Accept a root path in client metadata document URLs
Draft -02 section 3 marks a root path as NOT RECOMMENDED, not invalid, so
https://example.com/ is a valid client_id. A URL with no path is still
rejected.
* fix: Reject client metadata documents with private key material
A jwks key with private or symmetric members (d, p, q, dp, dq, qi, oth,
k) makes the document invalid, so it is not cached (draft -02 section 4.1).
Public keys are still ignored.
* fix: Respect no-cache, no-store and Age for client metadata documents
no-store and no-cache now expire the cached document at once, so it is
fetched again before the next authorization. Freshness from max-age or
Expires (measured from Date) is reduced by Age before clamping (draft -02
section 5.2).
* feat: Copy DCR and CIMD client logos to file storage in the background
logo_uri was rendered directly on the consent screen, which lets the logo
host track users and change the image at any time (draft -02 sections 8.8
and 9.2). A background task now fetches the logo with the strict fetch
rules, accepts only PNG, JPEG, GIF and WebP by their leading bytes, and
stores it content addressed in the public bucket. Requests never wait for
the upload.
* fix: Match loopback redirect URIs on any port
Native apps such as Claude Code register http://localhost/callback and
then listen on a free port, so an exact match rejected them. Loopback
redirect URIs now match on any port (RFC 8252 section 7.3); the scheme,
host and path must still match, and other URIs still match exactly.
* fix: Only break the redirect URL mid-word on the consent screen
break-all on the whole notice split words such as "continue" in the
loopback warning. It now applies only to the redirect URL.
* chore: Shorten the loopback warning on the consent screen
* fix: Show the punycode host of client metadata documents on the consent screen
A lookalike Unicode host was shown as written. Hostnames now come from a
shared getHostname helper that returns the punycode form.
* refactor: Move image type detection to ImageHelper and use it for avatars
Avatars and integration logos copied from a URL are now stored only if
they are PNG, JPEG, GIF or WebP, with the content type detected from the
file data rather than the remote host.
* fix: Address review findings for OAuth client logos and metadata caching
- Count the time since the Date header as age when caching metadata
documents, so a document is not cached past its Expires time
- Log logo scheduling failures instead of failing registration requests
- Run dynamic client registration in a transaction
- Delete copied logos from storage when no client uses them
* fix: Store one logo per OAuth client and return a distinct metadata document error
- Logos are stored per client, so a file is only deleted when its own
client stops using it. This removes the reference count and its race.
- A metadata document that cannot be fetched or is invalid now returns
invalid_client_metadata with the reason, not a 404. Network errors
are only logged.
- Move scheduleOptional to BaseTask so that any task can use it.
* fix: Allow private-use URI schemes for OAuth redirect URIs
Native apps receive OAuth redirects on a private-use URI scheme
(RFC 8252 section 7.1), e.g. com.example.app:/oauth2redirect or
myapp://callback. These were rejected when registering a client
because IsUrlList only accepted http, https and ftp, and the
recommended com.example.app:/path form was also rejected at authorize
time because it has no host.
Add isPrivateUseSchemeUrl, which accepts private-use schemes while
excluding web schemes and schemes that can execute or embed content,
and use it in both places.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix: Require a protocol for OAuth redirect URIs
isURL treats a bare word such as "callback" as a hostname, so a client
could be saved with a redirect URI that validateRedirectUri would always
reject at authorize time. Require a protocol when validating the list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix: Accept one-character private-use URI destinations
The destination check compared the pathname length, so a valid opaque URI
such as myapp:a was rejected along with the empty myapp: and myapp:/.
Reject only those empty forms.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* feat: Align MCP server and OAuth provider with newer MCP spec recommendations
- Upgrade @modelcontextprotocol/sdk to 1.31.0
- Include iss in authorization responses and advertise support (RFC 9207)
- Include required scopes in the WWW-Authenticate challenge on /mcp
- Expose title, website and icons in the MCP server info
- Centralise issuer, scopes, grant types and client type mapping in OAuthHelper
* Remove OAuth authorization issuer changes
* feat: Sync group descriptions and re-sync groups on SSO validation
- Sync group descriptions from group sync providers alongside names
- Lock name and description edits for externally synced groups at the
API layer and in the edit dialog
- Move group sync into a SyncUserGroupsTask, scheduled from login and
from ValidateSSOAccessTask after a fresh provider validation
- Add authenticationProviders.startGroupSync for providers that require
an admin consent flow before group sync can be enabled
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* fix: Make group sync setup generic and deduplicate sync jobs
- Expose groupSyncRequiresSetup from the provider list and use it in the
settings scene instead of checking for the azure provider name
- Rename the setup redirect query param to groupSync
- Give SyncUserGroupsTask a job id per user authentication so concurrent
syncs for the same authentication are not queued twice
* Use existing hook
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Adds a server-shared Vitest project with isolate disabled and moves every
server test into it by default. Files that mock or reset modules opt out
with a header marker and stay in the isolated project.
Restores env values that leaked between files: the test URL is reset after
each file, and tests that mutate env restore the previous value. The test
server no longer closes the shared sequelize pool.
* fix: Validate share id on shares.sitemap, shares.info, and shares.subscribe
Fixes OUTLINE-CLOUD-CWY
* test: Add regression tests for invalid share ids on shares.info and shares.subscribe
* test: Use an id that fails share slug validation and assert the message
* fix: batch large deletes in teamPermanentDeleter to shorten lock time
Delete events, collections, and documents in batches of ids outside the
final transaction so row locks are held briefly instead of for the whole
team deletion, which blocked the hourly old-events cleanup for hours.
* fix: type batched destroy helper structurally
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* fix: guard whole team deletion with a non-blocking mutex
Move the concurrency guard from the final transaction to the whole
routine so a second worker skips before running the batched deletes.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* checkpoint
* feat: Convert @ prefixed links to mentions on server Markdown writes
A Markdown link preceded by "@" now becomes a mention when brought in
through the documents API, MCP, or an import, so an exported mention
survives a round trip instead of degrading to a plain link.
The editor rule creates a generic mention carrying the href, and the
type of resource it points at is resolved by the plugin that recognizes
the URL through a new MentionProvider hook. Mentions of external
resources also serialize to their real URL rather than a mention:// URI,
so the reference resolves outside of Outline.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: Escape values interpolated into rendered mention HTML
markdown-it leaves markup in a link label intact when html is disabled,
so the label and attributes are escaped before being written into the
tag.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: Make count-based limit checks reliable under concurrency
Limits enforced by counting rows before an insert could be exceeded by
requests that ran at the same time, as each one read the same count
before any of them had written. Add a LockHelper that takes a Postgres
advisory lock for the remainder of the transaction, and use it to
serialize the check for webhook subscriptions, team domains, share
subscriptions and pins. The counts now also run on the transaction so
they include rows written earlier in the same request.
* Remove excessive tests
* Also guard simultaneous row removal
* fix: Normalize path when keying rate limiters
The router matches paths case-insensitively and with an optional trailing
slash, so those variants each received their own quota. Collapse them onto
a single key at both registration and consumption.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feedback
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>