Replace node-fetch v2 with Node 22's built-in fetch API across the entire server.
Add undici as an explicit dependency to provide the dispatcher API needed for
DNS pinning (SSRF rebinding prevention) in ssrfGuard.ts. All seven service files
that used a plain `import fetch from 'node-fetch'` are updated to use the global.
The ssrfGuard safeFetch/createPinnedAgent is rewritten as createPinnedDispatcher
using an undici Agent, with correct handling of the `all: true` lookup callback
required by Node 18+. The collabService dynamic require() and notifications agent
option are updated to use the dispatcher pattern. Test mocks are migrated from
vi.mock('node-fetch') to vi.stubGlobal('fetch'), and streaming test fixtures are
updated to use Web ReadableStream instead of Node Readable.
Fix several bugs in the Synology and Immich photo integrations:
- pipeAsset: guard against setting headers after stream has already started
- _getSynologySession: clear stale SID and re-login when decrypt_api_key returns null
instead of propagating success(null) downstream
- _requestSynologyApi: return retrySession error (not stale session) on retry failure;
also retry on error codes 106 (timeout) and 107 (duplicate login), not only 119
- searchSynologyPhotos: fix incorrect total field type (Synology list_item returns no
total); hasMore correctly uses allItems.length === limit
- _splitPackedSynologyId: validate cache_key format before use; callers return 400
- getImmichCredentials / _getSynologyCredentials: treat null from decrypt_api_key as
a missing-credentials condition rather than casting null to string
- Synology size param: enforce allowlist ['sm', 'm', 'xl'] per API documentation
Times like 09:00 were exported as YYYYMMDDTHHMM (13 chars) instead of
YYYYMMDDTHHMMSS (15 chars). Google Calendar couldn't parse the short
format and defaulted all events to 12:00 AM. Closes#432
- Add quantity field to packing items (persisted, visible per item)
- Bags are now renamable (click to edit in sidebar)
- Bags support multiple user assignments with avatar display
- New packing_bag_members table for multi-user bag ownership
- Save current packing list as reusable template
- Add bag members API endpoint (PUT /bags/:bagId/members)
- Migration 74: quantity on packing_items, user_id on packing_bags, packing_bag_members table
- Link budget items to reservations via reservation_id column
- Update budget entry when reservation price changes (not create duplicate)
- Delete budget entry when reservation price is cleared
- Sync price back to reservation when edited in budget panel
- Lock budget item name when linked to a reservation
- Add migration 73 for reservation_id on budget_items
- TRIP-002 now expects null dates and 7 placeholder days instead of forced date window
- Add TRIP-002b to verify custom day_count creates correct number of days
- Show day count input in trip form when no start/end date is set
- Backend accepts day_count param for create and update
- Remove forced date assignment for dateless trips (was always setting tomorrow + 7)
- Fix off-by-one: single-date fallback now creates 7 days instead of 8
- Add dayCount/dayCountHint translations for all 13 languages
GPX files containing both <wpt> and <trk> elements would only import
waypoints, silently discarding track geometry. The fallback chain only
parsed <trkpt> when no waypoints were found.
Replaced all regex-based XML parsing helpers with fast-xml-parser for
correctness (namespaces, CDATA, attribute ordering). Tracks are now
always parsed independently of waypoints, with each <trk> element
becoming its own place with route geometry. Fixes#427.
The search_place MCP tool was hardcoding a direct Nominatim call, ignoring
any configured Google Maps API key and never returning google_place_id despite
the tool description advertising it. Replace the inline fetch with the existing
searchPlaces() service which already switches between Google and Nominatim.
Update unit tests to mock mapsService instead of global fetch, and add a
dedicated test case for the Google path returning google_place_id.
Closes#424
Raw avatar filename was passed through formatFile without being
transformed into a full URL path, causing the browser to resolve
it relative to the current /trips/... page. Closes#417.
- Fix createAlbumLink using old column name (immich_album_id → album_id)
- Fix deleteAlbumLink not removing associated photos (with owner check)
- Update integration tests for new schema (asset_id, album_id, provider)
Encrypted webhook URLs are no longer returned to the frontend. Both user
and admin webhook fields now show '••••••••' as a placeholder when a URL
is already saved, and the sentinel value is skipped on save/test so the
stored secret is never exposed or accidentally overwritten.
Node 20+ enables autoSelectFamily by default, causing internal dns lookups
to be called with `all: true`. This expects the callback to receive an array
of address objects instead of a flat (address, family) pair, causing webhook
requests to fail with "Invalid IP address: undefined".
- migrations.ts: keep dev's migrations 69 (place_regions) + 70 (visited_regions), renumber our notification_channel_preferences migration to 71 and drop-old-table to 72
- translations: use dev values for existing keys, add notification system keys unique to this branch
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- SSRF: guard sendWebhook() with checkSsrf() + createPinnedAgent() to block
requests to loopback, link-local, private network, and cloud metadata endpoints
- XSS: escape subject, body, and ctaHref in buildEmailHtml() via escapeHtml()
to prevent HTML injection through user-controlled params (actor, preview, etc.)
- Encrypt webhook URLs at rest: apply maybe_encrypt_api_key on save
(settingsService for user URLs, authService for admin URL) and decrypt_api_key
on read in getUserWebhookUrl() / getAdminWebhookUrl()
- Log failed channel dispatches: inspect Promise.allSettled() results and log
rejections via logError instead of silently dropping them
- Log admin webhook failures: replace fire-and-forget .catch(() => {}) with
.catch(err => logError(...)) and await the call
- Migration 69: guard against missing notification_preferences table on fresh installs
- Migration 70: drop the now-unused notification_preferences table
- Refactor: extract applyUserChannelPrefs() helper to deduplicate
setPreferences / setAdminPreferences logic
- Tests: add SEC-016 (XSS, 5 cases) and SEC-017 (SSRF, 6 cases) test suites;
mock ssrfGuard in notificationService tests
- Add visited_regions table migration
- Mark/unmark region endpoints with auto-mark parent country
- Unmark country cascades to its regions; unmark last region cascades to country
- Region modal with mark/unmark flow and bucket list shortcut
- Viewport-based lazy loading of region GeoJSON at zoom >= 6
- i18n: add atlas.markRegionVisitedHint and atlas.confirmUnmarkRegion across all 13 locales
- Zoom >= 5 shows visited regions (states/provinces/departments) colored on the map
- Server resolves places to regions via Nominatim reverse geocoding (zoom=8)
- Supports all ISO levels: lvl4 (states), lvl5 (provinces), lvl6 (departments)
- Handles city-states (Berlin, Vienna, Hamburg) via city/county fallback
- Fuzzy name matching between Nominatim and GeoJSON for cross-format compatibility
- 10m admin_1 GeoJSON loaded server-side (cached), filtered per country
- Region colors match their parent country color
- Custom DOM tooltip (ref-based, no re-renders on hover)
- Country layer dims to 35% opacity when regions visible
- place_regions DB table caches resolved regions permanently
- Rate-limited Nominatim calls (1 req/sec) with progressive resolution
Introduces a fully featured notification system with three delivery
channels (in-app, email, webhook), normalized per-user/per-event/
per-channel preferences, admin-scoped notifications, scheduled trip
reminders and version update alerts.
- New notificationService.send() as the single orchestration entry point
- In-app notifications with simple/boolean/navigate types and WebSocket push
- Per-user preference matrix with normalized notification_channel_preferences table
- Admin notification preferences stored globally in app_settings
- Migration 69 normalizes legacy notification_preferences table
- Scheduler hooks for daily trip reminders and version checks
- DevNotificationsPanel for testing in dev mode
- All new tests passing, covering dispatch, preferences, migration, boolean
responses, resilience, and full API integration (NSVC, NPREF, INOTIF,
MIGR, VNOTIF, NROUTE series)
- Previous tests passing
- Budget category uses dropdown with existing categories instead of freetext
- Auto category uses translated booking type names (e.g. "Volo" in Italian)
- Remove number input spinner arrows, use decimal inputMode
- Add budget entry creation to PUT handler (update), not just POST (create)
- Error logging for failed budget entry creation
- i18n keys for all 13 languages
- Optional price and budget category fields on the reservation form
- When a price is set, a budget entry is automatically created on save
- Price and category stored in reservation metadata for reference
- Hint text shown when price is entered
- i18n keys for EN and DE
Replace inline DB queries in getTripSummary with calls to existing
service functions: listDays, listAccommodations, listBudgetItems,
listPackingItems, listReservations, listCollabNotes, getTripOwner,
and listMembers.
Budget and packing stats are now derived from the service results
instead of separate COUNT/SUM queries.
Replace all db.prepare() calls in mcp/index.ts, mcp/resources.ts, and
mcp/tools.ts with calls to the service layer. Add missing service functions:
- authService: isDemoUser, verifyMcpToken, verifyJwtToken
- adminService: isAddonEnabled
- atlasService: listVisitedCountries
- tripService: getTripSummary, listTrips with null archived param
Also fix getAssignmentWithPlace and formatAssignmentWithPlace to expose
place_id, assignment_time, and assignment_end_time at the top level, and
fix updateDay to correctly handle null title for clearing.
Add comprehensive unit and integration test suite for the MCP layer (821 tests all passing).
Dateless days are now reassigned to the new date range instead of being
deleted and recreated. This keeps all assignments, notes, bookings and
other day content intact when a user adds start/end dates to a trip
that was created without them.
- New todo_items DB table with priority, due date, description, user assignment
- Full CRUD API with WebSocket real-time sync
- 3-column UI: sidebar filters (All, My Tasks, Overdue, Done, by Priority),
task list with inline badges, and detail/create pane
- Apple-inspired design with custom dropdowns, date picker, priority system (P1-P3)
- Mobile responsive: icon-only sidebar, bottom-sheet modals for detail/create
- Lists tab with sub-tabs (Packing List + To-Do), persisted selection
- Addon renamed from "Packing List" to "Lists"
- i18n keys for all 13 languages
- UI polish: notification colors use system theme, mobile navbar cleanup,
settings page responsive buttons
When unlinking an Immich album, photos synced from that album are now
deleted. A new `album_link_id` FK column on `trip_photos` tracks the
source album link at sync time; `deleteAlbumLink` deletes matching
photos before removing the link. Individually-added photos are
unaffected. The client now refreshes the photo grid after unlinking.
Adds integration tests IMMICH-020 through IMMICH-024.
Closes#398
Now that trips always default to a start+7 day window when no dates
are provided, the test expectation of null dates and zero dated days
is no longer valid.
canAccessUserPhoto was using .get() which only returned the first matching
trip, causing access to be incorrectly denied when a photo was shared across
multiple trips and the requester was a member of a non-first trip.
Adds optional check_in and check_out fields to create_reservation and
link_hotel_accommodation so MCP clients can set accommodation times,
matching the existing REST API behaviour.
Closes#363
- No dates → tomorrow to tomorrow+7d
- Start only → end = start+7d
- End only → start = end-7d
- Both provided → unchanged
fix(ci): include client/package-lock.json in version bump commit
Trip members viewing another member's shared photo were getting a 404
because the proxy endpoints always used the requesting user's Immich
credentials instead of the photo owner's. The ?userId= query param the
client already sent was silently ignored.
- Add canAccessUserPhoto() to verify the asset is shared and the
requesting user is a trip member before allowing cross-user proxying
- Pass optional ownerUserId through proxyThumbnail, proxyOriginal, and
getAssetInfo so credentials are fetched for the correct user
- Enforce shared=1 check so unshared photos remain inaccessible
- toggleCompanyHoliday now calls loadStats() so the entitlement sidebar
updates immediately when a vacation day is converted to a company holiday
- deleteYear now deletes vacay_user_years rows for the removed year,
preventing stale entitlement data from persisting and re-appearing
when the year is re-created
- deleteYear recalculates carry-over for year+1 when year N is deleted,
using the new actual previous year as the source
- removeYear store action now calls loadStats() so the sidebar reflects
the recalculated carry-over without requiring a page refresh
- Add prev-year button (+[<] 2026 [>]+) so users can add years going
backwards after deleting a past year; add vacay.addPrevYear i18n key
to all 13 supported languages
Closes#371
When a user enters an http:// Immich URL that redirects to https://,
the test succeeded (GET follows redirects fine) but subsequent POST
requests (e.g. photo search) broke due to method downgrade on 301/302.
Now testConnection() checks resp.url against the input URL after a
successful fetch. If the only difference is http→https on the same
host and port, it returns a canonicalUrl so the frontend can update
the input field before the user saves — ensuring the correct URL is
stored.
- Fix attachment URLs to use /api/trips/:id/files/:id/download instead
of /uploads/files/... which was unconditionally blocked with 401
- Use getAuthUrl() with ephemeral tokens for displaying attachments and
opening them in a new tab (images, PDFs, documents)
- Replace htmlFor/id label pattern with ref.current.click() for the
file picker button in NoteFormModal — fixes file not being added to
pending list on first note creation
- Add integration tests COLLAB-028 to COLLAB-031 covering URL format,
listing URLs, ephemeral token download, and unauthenticated 401
* add test suite, mostly covers integration testing, tests are only backend side
* workflow runs the correct script
* workflow runs the correct script
* workflow runs the correct script
* unit tests incoming
* Fix multer silent rejections and error handler info leak
- Revert cb(null, false) to cb(new Error(...)) in auth.ts, collab.ts,
and files.ts so invalid uploads return an error instead of silently
dropping the file
- Error handler in app.ts now always returns 500 / "Internal server
error" instead of forwarding err.message to the client
* Use statusCode consistently for multer errors and error handler
- Error handler in app.ts reads err.statusCode to forward the correct
HTTP status while keeping the response body generic
New POST /api/trips/:id/copy endpoint that deep copies all trip
planning data (days, places, assignments, reservations, budget,
packing, accommodations, day notes) with proper FK remapping
inside a transaction. Skips files, collab data, and members.
Copy button on all dashboard card types (spotlight, grid, list,
archived) gated by trip_create permission. Translations for all
12 languages.
Also adds reminder_days to Trip interface (removes as-any casts).
Introduces a full in-app notification system with three types (simple,
boolean with server-side callbacks, navigate), three scopes (user, trip,
admin), fan-out persistence per recipient, and real-time push via
WebSocket. Includes a notification bell in the navbar, dropdown, dedicated
/notifications page, and a dev-only admin tab for testing all notification
variants.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Fixes#306 — OIDC scopes were hardcoded to 'openid email profile',
causing OIDC_ADMIN_CLAIM-based role mapping to fail when the required
scope (e.g. 'groups') wasn't requested. The new OIDC_SCOPE variable
defaults to 'openid email profile groups' so group-based admin mapping
works out of the box. Variable is now documented in README, docker-compose,
.env.example, and the Helm chart values.
Allow the first-boot admin account to be configured via ADMIN_EMAIL and
ADMIN_PASSWORD environment variables. If both are set the account is created
with those credentials; otherwise the existing random-password fallback is
used. Documented across .env.example, docker-compose.yml, Helm chart
(values.yaml, secret.yaml, deployment.yaml), and CLAUDE.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds COOKIE_SECURE (fixes login loop on plain-HTTP setups) and the previously
undocumented OIDC_DISCOVERY_URL to .env.example, docker-compose.yml, README.md,
chart/values.yaml, chart/templates/configmap.yaml, and chart/README.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Auth middleware now tags its 401s with code: AUTH_REQUIRED so the
client interceptor only redirects to /login on genuine session failures,
not on upstream API errors
- Fix /albums and album sync routes using raw encrypted API key instead
of getImmichCredentials() (which decrypts it), causing Immich to reject
requests with 401
- Add toast error notifications for all Immich operations in MemoriesPanel
that previously swallowed errors silently
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
db was already imported as addonDb; the extra db named import was
unnecessary. Updated the one stray db.prepare call at line 155 to use
addonDb consistently.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- LoginPage now uses getApiErrorMessage() instead of err.message so
backend validation errors (e.g. "Password must be at least 8 characters")
are displayed instead of the generic "Request failed with status code 400"
- Add missing db import in server/src/index.ts
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
security: require auth for uploaded photos (GHSA-wxx3-84fc-mrx2)
GHSA-pcr3-6647-jh72 (HIGH):
- Add canAccessTrip check to all /trips/:tripId/photos and
/trips/:tripId/album-links endpoints
- Prevents authenticated users from accessing other trips' photos
GHSA-wxx3-84fc-mrx2 (LOW):
- /uploads/photos now requires JWT auth token or valid share token
- Covers and avatars remain public (needed for login/share pages)
- Files were already blocked behind auth
Map markers:
- Collapsing a day in the sidebar hides its places from the map
- Places assigned to multiple days only hide when all days collapsed
- Unplanned places always stay visible
Immich settings:
- New POST /integrations/immich/test endpoint validates credentials
without saving them
- Save button disabled until test connection passes
- Changing URL or API key resets test status
- i18n: testFirst key for all 12 languages
- Link Immich albums to trips — photos sync automatically
- Album picker shows all user's Immich albums
- Linked albums displayed as chips with sync/unlink buttons
- Auto-sync on link: fetches all album photos and adds to trip
- Manual re-sync button for each linked album
- DB migration: trip_album_links table
fix: shared Immich photos visible to other trip members
- Thumbnail/original proxy now uses photo owner's Immich credentials
when userId query param is provided, fixing 404 for shared photos
- i18n: album keys for all 12 languages
Store & re-render optimization:
- TripPlannerPage uses selective Zustand selectors instead of full store
- placesSlice only updates affected days on place update/delete
- Route calculation only reacts to selected day's assignments
- DayPlanSidebar uses stable action refs instead of full store
Map marker performance:
- Shared photoService for PlaceAvatar and MapView (single cache, no duplicate requests)
- Client-side base64 thumbnail generation via canvas (CORS-safe for Wikimedia)
- Map markers use base64 data URL <img> tags for smooth zoom (no external image decode)
- Sidebar uses same base64 thumbnails with IntersectionObserver for visible-first loading
- Icon cache prevents duplicate L.divIcon creation
- MarkerClusterGroup with animate:false and optimized chunk settings
- Photo fetch deduplication and batched state updates
Server optimizations:
- Wikimedia image size reduced to 400px (from 600px)
- Photo cache: 5min TTL for errors (was 12h), prevents stale 404 caching
- Removed unused image-proxy endpoint
UX improvements:
- Splash screen with plane animation during initial photo preload
- Markdown rendering in DayPlanSidebar place descriptions
- Missing i18n keys added, all 12 languages synced to 1376 keys
- New expense_date column on budget items (DB migration #42)
- Date column in budget table with custom date picker
- CSV export button with BOM, semicolon separator, localized dates,
currency in header, per-person/day calculations
- CustomDatePicker compact/borderless modes for inline table use
- i18n keys for all 12 languages
Eliminates XSS token theft risk by storing session JWTs in an httpOnly
cookie (trek_session) instead of localStorage, making them inaccessible
to JavaScript entirely.
- Add cookie-parser middleware and setAuthCookie/clearAuthCookie helpers
- Set trek_session cookie on login, register, demo-login, MFA verify, OIDC exchange
- Auth middleware reads cookie first, falls back to Authorization: Bearer (MCP unchanged)
- Add POST /api/auth/logout to clear the cookie server-side
- Remove all localStorage auth_token reads/writes from client
- Axios uses withCredentials; raw fetch calls use credentials: include
- WebSocket ws-token exchange uses credentials: include (no JWT param)
- authStore initialises isLoading: true so ProtectedRoute waits for /api/auth/me
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Import places from shared Google Maps lists via URL.
Button in places sidebar next to GPX import opens a modal
where users can paste a shared list link. Server fetches
list data from Google Maps and creates places with name,
coordinates and notes. i18n keys added for all 12 languages.
Closes#205
Previously, when the JWT secret was used as a fallback encryption key,
nothing was written to data/.encryption_key. This meant that rotating
the JWT secret via the admin panel would silently break decryption of
all stored secrets on the next restart.
Now, whatever key is resolved — env var, JWT secret fallback, or
auto-generated — is immediately persisted to data/.encryption_key.
On all subsequent starts, the file is read directly and the fallback
chain is skipped entirely, making JWT rotation permanently safe.
The env var path also writes to the file so the key survives container
restarts if the env var is later removed.
process.exit(1) when ENCRYPTION_KEY is unset was a breaking change for
existing installs — a plain git pull would prevent the server from
starting.
Replace with a three-step fallback:
1. ENCRYPTION_KEY env var (explicit, takes priority)
2. data/.jwt_secret (existing installs: encrypted data stays readable
after upgrade with zero manual intervention)
3. data/.encryption_key auto-generated on first start (fresh installs)
A warning is logged when falling back to the JWT secret so operators
are nudged toward setting ENCRYPTION_KEY explicitly.
Update README env table and Docker Compose comment to reflect that
ENCRYPTION_KEY is recommended but no longer required.
Add server/scripts/migrate-encryption.ts — a standalone script that
re-encrypts all at-rest secrets (OIDC client secret, SMTP password,
Maps/OpenWeather/Immich API keys, MFA secrets) when rotating
ENCRYPTION_KEY, without requiring the app to be running.
- Prompts for old and new keys interactively; input is never echoed,
handles copy-pasted keys correctly via a shared readline interface
with a line queue to prevent race conditions on piped/pasted input
- Creates a timestamped DB backup before any changes
- Idempotent: detects already-migrated values by trying the new key
- Exits non-zero and retains the backup if any field fails
README updates:
- Add .env setup step (openssl rand -hex 32) before the Docker Compose
snippet so ENCRYPTION_KEY is set before first start
- Add ENCRYPTION_KEY to the docker run one-liner
- Add "Rotating the Encryption Key" section documenting the script,
the docker exec command, and the upgrade path via ./data/.jwt_secret
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The startup error now tells operators exactly where to find the old key
value (./data/.jwt_secret) rather than just saying "your old JWT_SECRET".
docker-compose.yml and README updated to mark ENCRYPTION_KEY as required
and remove the stale "auto-generated" comments.
Auto-generating and persisting the key to data/.encryption_key co-locates
the key with the database, defeating encryption at rest if an attacker can
read the data directory. It also silently loses all encrypted secrets if the
data volume is recreated.
Replace the auto-generation fallback with a hard startup error that tells
operators exactly what to do:
- Upgraders from the JWT_SECRET-derived encryption era: set ENCRYPTION_KEY
to their old JWT_SECRET so existing ciphertext remains readable.
- Fresh installs: generate a key with `openssl rand -hex 32`.
- Create server/src/utils/ssrfGuard.ts with checkSsrf() and createPinnedAgent()
- Resolves DNS before allowing outbound requests to catch hostnames that
map to private IPs (closes the TOCTOU gap in the old inline checks)
- Always blocks loopback (127.x, ::1) and link-local/metadata (169.254.x)
- RFC-1918, CGNAT (100.64/10), and IPv6 ULA ranges blocked by default;
opt-in via ALLOW_INTERNAL_NETWORK=true for self-hosters running Immich
on a local network
- createPinnedAgent() pins node-fetch to the validated IP, preventing
DNS rebinding between the check and the actual connection
- Replace isValidImmichUrl() (hostname-string check, no DNS resolution)
with checkSsrf(); make PUT /integrations/immich/settings async
- Audit log entry (immich.private_ip_configured) written when a user
saves an Immich URL that resolves to a private IP
- Response includes a warning field surfaced as a toast in the UI
- Replace ~20 lines of duplicated inline SSRF logic in the link-preview
handler with a single checkSsrf() call + pinned agent
- Document ALLOW_INTERNAL_NETWORK in README, docker-compose.yml,
server/.env.example, chart/values.yaml, chart/templates/configmap.yaml,
and chart/README.md
Replace duplicated inline validation with a shared validatePassword()
utility that checks minimum length (8), rejects repetitive and common
passwords, and requires uppercase, lowercase, a digit, and a special
character.
- Add server/src/services/passwordPolicy.ts as single source of truth
- Apply to registration, password change, and admin create/edit user
(admin routes previously had zero validation)
- Fix client min-length mismatch (6 vs 8) in RegisterPage and LoginPage
- Add client-side password length guard to AdminPage forms
- Update register.passwordTooShort and settings.passwordWeak i18n keys
in all 12 locales to reflect the corrected requirements
TOTP brute-force is a realistic attack once a password is compromised:
with no independent throttle, an attacker shared the login budget (10
attempts) across /login, /register, and /mfa/verify-login, and
/mfa/enable had no rate limiting at all.
- Add a dedicated `mfaAttempts` store so MFA limits are tracked
separately from login attempts
- Introduce `mfaLimiter` (5 attempts / 15 min) applied to both
/mfa/verify-login and /mfa/enable
- Refactor `rateLimiter()` to accept an optional store parameter,
keeping all existing call-sites unchanged
- Include mfaAttempts in the periodic cleanup interval
Three vulnerabilities patched in the /export.ics route:
- esc() now handles bare \r and CRLF sequences — the previous regex only
matched \n, leaving \r intact and allowing CRLF injection via \r\n
- reservation DESCRIPTION field was built from unescaped user data
(type, confirmation_number, notes, airline, flight/train numbers,
airports) and written raw into ICS output; now passed through esc()
- Content-Disposition filename used ICS escaping instead of HTTP header
sanitization; replaced with a character allowlist to prevent " and
\r\n injection into the response header
Setting JWT_SECRET via environment variable was broken by design:
the admin panel rotation updates the in-memory binding and persists
the new value to data/.jwt_secret, but an env var would silently
override it on the next restart, reverting the rotation.
The server now always loads JWT_SECRET from data/.jwt_secret
(auto-generating it on first start), making the file the single
source of truth. Rotation is handled exclusively through the admin
panel.
- config.ts: drop process.env.JWT_SECRET fallback and
JWT_SECRET_IS_GENERATED export; always read from / write to
data/.jwt_secret
- index.ts: remove the now-obsolete JWT_SECRET startup warning
- .env.example, docker-compose.yml, README: remove JWT_SECRET entries
- Helm chart: remove JWT_SECRET from secretEnv, secret.yaml, and
deployment.yaml; rename generateJwtSecret → generateEncryptionKey
and update NOTES.txt and README accordingly
Introduces a dedicated ENCRYPTION_KEY for encrypting stored secrets
(API keys, MFA TOTP, SMTP password, OIDC client secret) so that
rotating the JWT signing secret no longer invalidates encrypted data,
and a compromised JWT_SECRET no longer exposes stored credentials.
- server/src/config.ts: add ENCRYPTION_KEY (auto-generated to
data/.encryption_key if not set, same pattern as JWT_SECRET);
switch JWT_SECRET to `export let` so updateJwtSecret() keeps the
CJS module binding live for all importers without restart
- apiKeyCrypto.ts, mfaCrypto.ts: derive encryption keys from
ENCRYPTION_KEY instead of JWT_SECRET
- admin POST /rotate-jwt-secret: generates a new 32-byte hex secret,
persists it to data/.jwt_secret, updates the live in-process binding
via updateJwtSecret(), and writes an audit log entry
- Admin panel (Settings → Danger Zone): "Rotate JWT Secret" button
with a confirmation modal warning that all sessions will be
invalidated; on success the acting admin is logged out immediately
- docker-compose.yml, .env.example, README, Helm chart (values.yaml,
secret.yaml, deployment.yaml, NOTES.txt, README): document
ENCRYPTION_KEY and its upgrade migration path
Before swapping in a restored database, run PRAGMA integrity_check and
verify the five core TREK tables (users, trips, trip_members, places,
days) are present. This blocks restoring corrupt, empty, or unrelated
SQLite files that would otherwise crash the app immediately after swap,
and prevents a malicious admin from hot-swapping a crafted database with
forged users or permissions.
Per-user Immich API keys were stored as plaintext in the users table,
giving any attacker with DB read access full control over each user's
Immich photo server. Keys are now encrypted on write with
maybe_encrypt_api_key() and decrypted at the point of use via a shared
getImmichCredentials() helper. A new migration (index 66) back-fills
encryption for any existing plaintext values on startup.
Addresses CWE-598: long-lived JWTs were exposed in WebSocket URLs, file
download links, and Immich asset proxy URLs, leaking into server logs,
browser history, and Referer headers.
- Add ephemeralTokens service: in-memory single-use tokens with per-purpose
TTLs (ws=30s, download/immich=60s), max 10k entries, periodic cleanup
- Add POST /api/auth/ws-token and POST /api/auth/resource-token endpoints
- WebSocket auth now consumes an ephemeral token instead of verifying the JWT
directly from the URL; client fetches a fresh token before each connect
- File download ?token= query param now accepts ephemeral tokens; Bearer
header path continues to accept JWTs for programmatic access
- Immich asset proxy replaces authFromQuery JWT injection with ephemeral token
consumption
- Client: new getAuthUrl() utility, AuthedImg/ImmichImg components, and async
onClick handlers replace the synchronous authUrl() pattern throughout
FileManager, PlaceInspector, and MemoriesPanel
- Add OIDC_DISCOVERY_URL env var and oidc_discovery_url DB setting to allow
overriding the auto-constructed discovery endpoint (required for Authentik
and similar providers); exposed in the admin UI and .env.example
The OIDC login route silently fell back to building the redirect URI from
X-Forwarded-Host/X-Forwarded-Proto when APP_URL was not configured. An
attacker could set X-Forwarded-Host: attacker.example.com to redirect the
authorization code to their own server after the user authenticates.
Remove the header-derived fallback entirely. If APP_URL is not set (via env
or the app_url DB setting), the OIDC login endpoint now returns a 500 error
rather than trusting attacker-controlled request headers. Document APP_URL
in .env.example as required for OIDC use.
The smtp_pass setting was stored as plaintext in app_settings, exposing
SMTP credentials to anyone with database read access. Apply the same
encrypt_api_key/decrypt_api_key pattern already used for OIDC client
secrets and API keys. A new migration transparently re-encrypts any
existing plaintext value on startup; decrypt_api_key handles legacy
plaintext gracefully so in-flight reads remain safe during upgrade.
The oidc_client_secret was written to app_settings as plaintext,
unlike Maps and OpenWeather API keys which are protected with
apiKeyCrypto. An attacker with read access to the SQLite file
(e.g. via a backup download) could obtain the secret and
impersonate the application with the identity provider.
- Encrypt on write in PUT /api/admin/oidc via maybe_encrypt_api_key
- Decrypt on read in GET /api/admin/oidc and in getOidcConfig()
(oidc.ts) before passing the secret to the OIDC client library
- Add a startup migration that encrypts any existing plaintext value
already present in the database
getPlaceWithTags, canAccessTrip, and isOwner were calling _db! directly,
bypassing the Proxy that guards against null-dereference during a backup
restore. When the restore handler briefly sets _db = null, any concurrent
request hitting these helpers would crash with an unhandled TypeError
instead of receiving a clean 503-style error.
Replace all four _db! accesses with the exported db proxy so the guard
("Database connection is not available") fires consistently.
Previously, the migration runner called each migration function directly with no transaction wrapping and updated schema_version only after all pending migrations had run. A mid-migration failure (e.g. disk full after ALTER TABLE but before CREATE INDEX) would leave the schema in a partially-applied state with no rollback path. On the next restart the broken migration would be skipped — because schema_version had not advanced — but only if the failure was noticed at all.
~43 catch {} blocks silently discarded every error, including non-idempotency errors such as disk-full or corruption, making it impossible to know a migration had failed.
Changes:
- Each migration now runs inside db.transaction(); better-sqlite3 rolls back automatically on throw.
- schema_version is updated after every individual migration succeeds, so a failure does not cause already-applied migrations to re-run.
- A migration that throws after rollback logs FATAL and calls process.exit(1), refusing to start with a broken schema.
- All catch {} blocks on ALTER TABLE ADD COLUMN re-throw any error that is not "duplicate column name", so only the expected idempotency case is swallowed.
- Genuinely optional steps (INSERT OR IGNORE, UPDATE data-copy, DROP TABLE IF EXISTS) now log a warning instead of discarding the error entirely.
The POST /api/admin/update endpoint ran git pull, npm install, and npm run build via execSync, potentially giving any compromised admin account full code execution on the host in case repository is compromised. TREK ships as a Docker image so runtime self-updating is unnecessary.
- Remove the /update route and child_process import from admin.ts
- Remove the installUpdate API client method
- Replace the live-update modal with an info-only panel showing docker pull instructions and a link to the GitHub release
- Drop the updating/updateResult state and handleInstallUpdate handler
Adds a full permissions management feature allowing admins to control
who can perform actions across the app (trip CRUD, files, places,
budget, packing, reservations, collab, members, share links).
- New server/src/services/permissions.ts: 16 configurable actions,
in-memory cache, checkPermission() helper, backwards-compatible
defaults matching upstream behaviour
- GET/PUT /admin/permissions endpoints; permissions loaded into
app-config response so clients have them on startup
- checkPermission() applied to all mutating route handlers across
10 server route files; getTripOwnerId() helper eliminates repeated
inline DB queries; trips.ts and files.ts now reuse canAccessTrip()
result to avoid redundant DB round-trips
- New client/src/store/permissionsStore.ts: Zustand store +
useCanDo() hook; TripOwnerContext type accepts both Trip and
DashboardTrip shapes without casting at call sites
- New client/src/components/Admin/PermissionsPanel.tsx: categorised
UI with per-action dropdowns, customised badge, save/reset
- AdminPage, DashboardPage, FileManager, PlacesSidebar,
TripMembersModal gated via useCanDo(); no prop drilling
- 46 perm.* translation keys added to all 12 language files
Bounding boxes overlap for neighboring countries (e.g. Munich matched
Austria instead of Germany). Now uses Nominatim reverse geocoding with
in-memory cache as primary fallback, bounding boxes only as last resort.
- Block direct access to /uploads/files (401), serve via authenticated
/api/trips/:tripId/files/:id/download with JWT verification
- Client passes auth token as query parameter for direct links
- Atlas country search now uses Intl.DisplayNames (user language) instead
of English GeoJSON names
- Atlas search results use flagcdn.com flag images instead of emoji
- Add configurable trip reminder days (1, 3, 9 or custom up to 30) settable by trip owner
- Grant administrators full access to edit, archive, delete, view and list all trips
- Show trip owner email in audit logs and docker logs when admin edits/deletes another user's trip
- Show target user email in audit logs when admin edits or deletes a user account
- Use email instead of username in all notifications (Discord/Slack/email) to avoid ambiguity
- Grey out notification event toggles when no SMTP/webhook is configured
- Grey out trip reminder selector when notifications are disabled
- Skip local admin account creation when OIDC_ONLY=true with OIDC configured
- Conditional scheduler logging: show disabled reason or active reminder count
- Log per-owner reminder creation/update in docker logs
- Demote 401/403 HTTP errors to DEBUG log level to reduce noise
- Hide edit/archive/delete buttons for non-owner invited users on trip cards
- Fix literal "0" rendering on trip cards from SQLite numeric is_owner field
- Add missing translation keys across all 14 language files
Made-with: Cursor
- Add centralized notification service with webhook (Discord/Slack) and
email (SMTP) support, triggered for trip invites, booking changes,
collab messages, and trip reminders
- Webhook sends one message per event (group channel); email sends
individually per trip member, excluding the actor
- Discord invite notifications now include the invited user's name
- Add LOG_LEVEL env var (info/debug) controlling console and file output
- INFO logs show user email, action, and IP for audit events; errors
for HTTP requests
- DEBUG logs show every request with full body/query (passwords redacted),
audit details, notification params, and webhook payloads
- Add persistent trek.log file logging with 10MB rotation (5 files)
in /app/data/logs/
- Color-coded log levels in Docker console output
- Timestamps without timezone name (user sets TZ via Docker)
- Add Test Webhook and Save buttons to admin notification settings
- Move notification event toggles to admin panel
- Add daily trip reminder scheduler (9 AM, timezone-aware)
- Wire up booking create/update/delete and collab message notifications
- Add i18n keys for notification UI across all 13 languages
Made-with: Cursor
- Add google_place_id and osm_id params to create_place tool so the app
can fetch opening hours and ratings for MCP-created places
- Add list_categories tool for discovering category IDs
- Add search_place tool (Nominatim) to look up osm_id before creating
