Add long-lived API tokens with three permission scopes (read-only,
deploy-only, full-admin) for CI/CD pipelines, scripts, and automation.
- Database: api_tokens table with SHA-256 hashed storage
- Auth: extend middleware to authenticate Bearer API tokens
- Scope enforcement: middleware restricts actions per token scope
- API: CRUD endpoints gated behind Team Pro + admin
- UI: ApiTokensSection in Settings Hub with create/revoke/copy flows
- Docs: new api-tokens.mdx with usage examples and screenshots
* feat: SSO & LDAP authentication for Team Pro
Add SSO integration allowing Team Pro users to authenticate via LDAP/Active Directory, Google, GitHub, and Okta identity providers. SSO works alongside password authentication with auto-provisioning and role mapping.
- LDAP bind+search authentication with group-based role mapping
- OIDC/OAuth2 flows with PKCE and CSRF protection for Google, GitHub, Okta
- Auto-provisioning: first SSO login creates a Sencho account automatically
- Role mapping via LDAP group membership or OIDC JWT claims
- SSO settings UI in Settings → SSO with per-provider config and test connection
- SSO login buttons on login page with LDAP toggle
- Environment variable seeding for infrastructure-as-code workflows
- Secrets encrypted at rest via CryptoService (AES-256-GCM)
- Seat limit enforcement during auto-provisioning
- Full documentation: feature docs, quickstart guides, env var reference
* fix: resolve ESLint errors in SSO feature
- Remove unnecessary escape characters in regex character classes
- Remove unused `issuer` variable from OIDC callback handler
- Fix setState-in-effect lint error in Login.tsx by using useState initializer
- Suppress set-state-in-effect for SSOSection fetch pattern (matches existing codebase convention)
* feat: audit logging, secrets at rest, and legacy cleanup
- Add Team Pro audit log: records all mutating API actions with user
attribution, searchable timeline UI with filtering and pagination
- Add AES-256-GCM encryption at rest for node API tokens via CryptoService
- Drop 9 legacy SSH/TLS columns from nodes table (dead since v0.7)
- Remove orphaned MaintenanceModal.tsx (dead code, never imported)
- Add requireTeamPro backend guard for team-tier features
- Add audit log cleanup (90-day retention) to MonitorService
- Add docs page and navigation entry for audit log feature
* fix: remove unused AUTH_TAG_LENGTH constant from CryptoService
* docs: align documentation theme with Sencho app branding
- Change dark background from #0F172A (blue-tinted slate) to #090909 (neutral dark)
- Update accent colors to match app brand oklch(0.72 0.14 200)
- Add custom.css to increase navbar and footer logo size from 28px to 36px
* fix(deps): resolve npm audit vulnerabilities in backend
Fix path-to-regexp (high) and brace-expansion (moderate) via npm audit fix.
* chore: remove CLAUDE.md and MANUAL_STEPS.md from git tracking
These files contain internal project instructions that should not be
public. Both are now in .gitignore to prevent re-tracking.
* chore: remove build output and generated files from git tracking
Remove backend/public/ (frontend build artifacts) and backend/cookies.txt
from tracking. Add demo-video/ and website/ to .gitignore.
* docs: align mintlify theme with website and app branding
Update colors to brand cyan, add Geist font, dark mode default,
gradient background, navbar with GitHub link and CTA, and footer
matching the website structure.
* feat: add RBAC viewer accounts, atomic deployments, and fleet-wide backups (Pro)
Introduces three Pro-tier features:
- RBAC: Multi-user system with admin/viewer roles, user management UI,
automatic migration from single-admin credentials, viewer restrictions
across the entire UI (read-only editor, hidden action buttons)
- Atomic Deployments: Pre-deploy file backup to .sencho-backup/, automatic
rollback on health probe failure, manual rollback button, health probes
added to stack updates, webhook-triggered deploys use atomic rollback
- Fleet-Wide Backups: Point-in-time snapshots of compose files across all
nodes (local + remote), stored centrally in SQLite, per-stack restore
with optional redeploy, graceful handling of offline nodes
* fix(settings): use correct ProGate prop name in UsersSection
* fix(settings): remove unused isPro prop from UsersSection
* fix(auth): fetch user info after login and setup so isAdmin is set correctly
* feat(pricing): revise pricing strategy and enforce variant-based seat limits
Raise Personal Pro from $49/yr to $69/yr with 3 viewer seats (up from 1).
Add $15/mo billing option for Team Pro. Mark lifetime pricing as a
90-day early-adopter offer. Store Lemon Squeezy variant_name on
activation/validation and enforce seat limits server-side per variant.
* feat(licensing): add Lemon Squeezy checkout, webhook, and billing portal integration
Server-side checkout URL generation (POST /api/checkout) with admin email
pre-fill and instance_id custom data. HMAC-SHA256 verified webhook endpoint
(POST /api/webhooks/lemonsqueezy) handling order, subscription, and payment
lifecycle events for automatic license activation. Customer billing portal
link stored from webhook events and exposed via GET /api/billing/portal.
In-app checkout buttons in Settings with manual license key fallback.
* fix(licensing): exempt Lemon Squeezy webhook from auth middleware
The catch-all auth middleware on /api/* was blocking the public webhook
endpoint. Added /webhooks/lemonsqueezy to the exemption list alongside
/auth/* and /webhooks/:id/trigger.
* feat(pricing): update pricing to final live rates
Personal Pro: $7.99/month, $69.99/year, $249 lifetime.
Team Pro: $49.99/month, $499.99/year, $1,499 lifetime.
Added personal_monthly checkout variant across backend, frontend, and website.
* refactor(licensing): remove server-side checkout/webhook for self-hosted model
Sencho is self-hosted — each user runs their own instance, so there is
no central server to receive webhooks or hold the store API key. Replaced
in-app checkout buttons with a "View Pricing" redirect to sencho.io and
kept manual license key activation as the primary flow.
- Delete LemonSqueezyService (checkout, webhook, HMAC verification)
- Remove POST /api/checkout, GET /api/billing/portal, POST /api/webhooks/lemonsqueezy
- Remove raw body parser and auth exemption for webhook route
- Remove all LEMONSQUEEZY_* env vars from .env.example
- Replace checkout buttons in SettingsModal with single "View Pricing" button
- Simplify LicenseContext checkout to open sencho.io pricing page
- Update licensing docs to reflect website-based purchase flow
* chore: normalize em-dashes to hyphens across codebase (linter)
* chore: remove accidentally tracked directories from index
* feat: add RBAC viewer accounts, atomic deployments, and fleet-wide backups (Pro)
Introduces three Pro-tier features:
- RBAC: Multi-user system with admin/viewer roles, user management UI,
automatic migration from single-admin credentials, viewer restrictions
across the entire UI (read-only editor, hidden action buttons)
- Atomic Deployments: Pre-deploy file backup to .sencho-backup/, automatic
rollback on health probe failure, manual rollback button, health probes
added to stack updates, webhook-triggered deploys use atomic rollback
- Fleet-Wide Backups: Point-in-time snapshots of compose files across all
nodes (local + remote), stored centrally in SQLite, per-stack restore
with optional redeploy, graceful handling of offline nodes
* fix(settings): use correct ProGate prop name in UsersSection
* fix(settings): remove unused isPro prop from UsersSection
* fix(auth): fetch user info after login and setup so isAdmin is set correctly
Add custom webhooks allowing external CI/CD systems (GitHub Actions, GitLab CI, etc.)
to trigger stack actions via HTTP POST with HMAC-SHA256 signature authentication.
Includes webhook CRUD management UI in Settings, execution history tracking,
one-time secret reveal, enable/disable toggle, and comprehensive documentation.
Add fleet health summary cards, container-level drill-down, node sorting
and filtering, fleet-wide search, and critical node detection to the
Fleet View. Fix ProGate to wrap placeholder content and add error toasts.
* feat(auth): redesign Login and Setup pages with split-panel branding layout
Replace plain Card-based Login and Setup forms with a professional
split-panel layout: always-dark branding panel (dot grid texture, logo,
tagline, cyan accent line) on the left, theme-aware form on the right.
Mobile collapses to single column with compact logo header.
- Add optional admin email field on Setup for license recovery
- Backend accepts and stores admin_email in setup endpoint
- Fix data-stacks-loaded attribute forwarding (wrap in div, use string values)
* docs: add license-settings and profile-dropdown screenshots
* fix(docs): correct logo paths from .svg to .png
* fix(deps): resolve picomatch high-severity vulnerability
Replace plain Card-based Login and Setup forms with a professional
split-panel layout: always-dark branding panel (dot grid texture, logo,
tagline, cyan accent line) on the left, theme-aware form on the right.
Mobile collapses to single column with compact logo header.
- Add optional admin email field on Setup for license recovery
- Backend accepts and stores admin_email in setup endpoint
- Fix data-stacks-loaded attribute forwarding (wrap in div, use string values)
* feat: add license gating system with Lemon Squeezy integration
Add Community/Pro tier infrastructure:
- LicenseService singleton with Lemon Squeezy license API integration
- /api/license endpoints (GET info, POST activate/deactivate/validate)
- 14-day Pro trial activated automatically on first boot
- 72-hour periodic validation with 30-day offline grace period
- LicenseContext provider for frontend tier awareness
- License settings tab with activation UI and status display
- ProBadge and ProGate reusable components for feature gating
- requirePro per-route guard for backend Pro-only endpoints
- Proxy bypass for /api/license routes (local-only, never proxied)
* feat: add user profile dropdown and reorganize top navigation
- Create UserProfileDropdown component with settings, billing, theme
toggle (System/Light/Dark), documentation links, and logout button
- Remove logout button from sidebar header
- Remove standalone settings button from top bar
- Move theme toggle from Settings modal to profile dropdown
- Inject app version via Vite define from root package.json
- Add globals.d.ts for __APP_VERSION__ type declaration
* refactor(settings): remove appearance tab from settings modal
Theme toggle was moved to the User Profile Dropdown in the previous
commit. Remove the now-redundant Appearance section, its nav button,
and the unused theme/setTheme props from SettingsModal.
* feat: add fleet view dashboard and about settings section
Fleet Overview: aggregates all nodes into a card grid showing status,
container counts, CPU/RAM/disk usage bars. Pro tier unlocks stack
drill-down with auto-refresh (30s). Backend endpoints /api/fleet/overview
and /api/fleet/node/:nodeId/stacks query nodes in parallel.
About section in Settings: displays version, license tier, status,
instance ID, and links to docs/changelog/issues.
Sidebar perf fix: stack status fetches now run in parallel via
Promise.allSettled instead of sequential for-loop, significantly
reducing load time for nodes with many stacks.
Also removes version number from User Profile Dropdown (now in About).
* fix(ci): resolve Docker build and E2E test failures
- Copy root package.json into frontend build stage so vite.config.ts
can read the app version during Docker multi-stage build.
- Update auth E2E test: logout button moved into User Profile Dropdown.
- Update nodes E2E test: Settings button moved into User Profile Dropdown.
Add https: to img-src CSP directive so App Store template icons from
external registries (raw.githubusercontent.com) load correctly. Disable
Origin-Agent-Cluster header (only meaningful over HTTPS). Add minWidth={0}
to Recharts ResponsiveContainer to suppress dimension warnings.
The /envs endpoint now filters to only return env files that actually exist
on disk, and absolute env_file paths from compose files (e.g. shared
globals.env in a sibling directory) are no longer rejected. The inline
theme-detection script is moved to an external file to comply with CSP.
* chore(deps): upgrade ESLint 9 → 10 with plugin compatibility fixes
- eslint: ^9 → ^10.0.0 (backend + frontend)
- @eslint/js: ^9 → ^10.0.0 (backend + frontend)
- eslint-plugin-react-refresh: ^0.4.24 → ^0.5.2 (ESM, config factory API)
- Update frontend eslint.config.js: destructured import for react-refresh,
call configs.vite() as factory function (0.5 API change)
- Downgrade new ESLint 10 rules (no-useless-assignment, preserve-caught-error)
to warnings for existing code patterns
- eslint-plugin-react-hooks stays at 7.0.1 (stable) with --legacy-peer-deps
until a stable release adds ESLint 10 peer support
* chore(deps): upgrade recharts 2.x to 3.8 with chart.tsx type fixes
recharts 3.x moved Tooltip/Legend props to context-based API.
Updated chart.tsx to use explicit prop interfaces with internal
recharts type imports (LegendPayload, TooltipPayload, TooltipFormatter).
* chore(deps): upgrade TypeScript 5.9 to 6.0
- Remove deprecated baseUrl from frontend tsconfig (paths works standalone in TS 6)
- Add react-is dependency required by recharts 3.x at build time
- Backend and frontend both compile and lint cleanly
* chore(deps): upgrade Vite 7.3 to 8.0 and @vitejs/plugin-react to 6.0
Vite 8 replaces Rollup+esbuild with Rolldown, significantly improving
build speed (~2s vs ~18s). No config changes required.
* fix(ci): add .npmrc with legacy-peer-deps for CI and Docker builds
typescript-eslint@8.x requires typescript <6.0.0 and
eslint-plugin-react-hooks@7.0.1 requires eslint <=9. Until upstream
packages release compatible versions, legacy-peer-deps is needed.
* docs: add logo assets and re-ignore CLAUDE.md
* fix(ci): copy .npmrc into prod-deps Docker stage
The prod-deps stage also runs npm ci with backend/package.json but was
missing the .npmrc needed to bypass peer dep conflicts.
Two console errors on HTTP deployments with no functional impact:
1. Helmet's default Cross-Origin-Opener-Policy: same-origin is ignored
by browsers over HTTP but logged as a console error. Disabled via
crossOriginOpenerPolicy: false (same rationale as HSTS/COEP).
2. Vite's production build injects an inline module-preload polyfill
script blocked by script-src 'self'. Disabled via
build.modulePreload.polyfill: false — all modern browsers support
link rel="modulepreload" natively.
Containers launched from the COMPOSE_DIR root (not individual subdirs)
all share the root folder's name as their com.docker.compose.project
label, making them appear as "external" despite being in COMPOSE_DIR.
Switch to com.docker.compose.project.working_dir: a container is
managed if its working directory is within COMPOSE_DIR, regardless
of what project name Docker Compose assigned to it.
Helmet 8 merges custom directives with its built-in defaults, which
include upgrade-insecure-requests. Simply omitting the directive from
the custom object (PR #59) was insufficient — Helmet silently re-adds
it from defaults. Setting upgradeInsecureRequests: null is the correct
Helmet 8 API to remove a default directive.
This was the root cause of the persistent blank page on plain-HTTP
self-hosted deployments: the directive tells browsers to upgrade all
HTTP sub-resource fetches to HTTPS, producing ERR_SSL_PROTOCOL_ERROR
on every JS/CSS asset.
- Add x-sencho-proxy sentinel header to all proxied responses so
the frontend can distinguish remote auth failures from local session
expiry, breaking the logout loop when a node's api_token expires
- Add authMiddleware to all 5 /api/notifications endpoints that were
missing protection (default-deny policy enforcement)
- Expand CSP to include connectSrc ws:/wss: and workerSrc blob:
for WebSocket and Monaco editor worker support
Helmet's defaults are designed for HTTPS-only deployments. Two directives
were actively breaking plain-HTTP self-hosted instances:
- upgrade-insecure-requests: causes browsers to upgrade all JS/CSS/asset
fetches to HTTPS → ERR_SSL_PROTOCOL_ERROR → completely blank page
- Strict-Transport-Security: permanently instructs browsers to refuse HTTP
for 1 year, even after the header is removed from the server
Also handle docker socket GID=0 (root:root) edge case in entrypoint and
add [entrypoint] diagnostic log lines for easier debugging.
- Classify all Docker images, volumes, and networks as managed (Sencho
stack), external (other Compose project), or unused/system via a new
getClassifiedResources() method and GET /api/system/resources endpoint
- Add pruneManagedOnly() + getDiskUsageClassified() to DockerController
- Prune buttons now default to Sencho-managed scope; "All Docker" is
hidden in a ⋮ dropdown with a distinct destructive confirm dialog
- Replace Reclaimable Space donut with interactive Docker Disk Footprint
widget (stacked bar with clickable segments that filter resource tabs)
- Add managed/external filter toggles and classification badges per tab
- GET /api/stats now returns managed + unmanaged container counts; Home
Dashboard Active Containers card subtitle shows "N managed · N external"
- Rename "Ghost Containers" tab/copy to "Unmanaged Containers" throughout
- Split serial build-and-test job into parallel backend, frontend, docker-validate, and e2e jobs
- backend job: build → test → lint (new ESLint) → npm audit --audit-level=high
- frontend job: build → lint → npm audit --audit-level=high
- docker-validate job: builds image on every PR without pushing; Trivy scans for CRITICAL/HIGH CVEs (informational)
- e2e job: runs full Playwright suite against live dev servers after backend+frontend pass
- Add backend/eslint.config.mjs and lint script to backend/package.json
- Fix rate limiter to allow 100 attempts in dev mode so E2E tests are
not blocked by failed-login attempts during test development
- Simplify logout button selector to use Lucide icon class (lucide-log-out)
instead of the fragile Tooltip-content locator chain that broke on navigation
- Rewrite stacks E2E spec: use waitForFunction to wait for sidebar to load,
delete leftover stacks via browser-context fetch before creating, and use
page.reload() to get a clean sidebar state
- Fix AlertDialogContent: remove asChild+motion.div pattern that triggered a
React.Children.only crash — Radix AlertDialog.Content injects a second
DescriptionWarning child internally, breaking Slot when asChild=true; replace
with CSS keyframe animations (data-[state=open]:animate-in)
- Fix final assertion in delete test to use exact text + listbox scope to avoid
false positives from similarly-named stacks like e2e-test-stack-*
- All 6 E2E tests pass (4 auth + 2 stacks); node tests skip gracefully
SECURITY (critical fixes):
- Add authMiddleware to /api/system/console-token (was publicly accessible)
- Validate api_url on node create/update to prevent SSRF (rejects localhost/loopback)
- Add rate limiting (5 req/15 min/IP) to /api/auth/login and /api/auth/setup
- Fix path traversal in env_file resolution — absolute/escaping paths rejected
- Add stack name validation to GET routes (was only on PUT/POST)
- Add helmet security headers middleware
- Restrict CORS to FRONTEND_URL in production
PRODUCTION READINESS:
- Add GET /api/health public endpoint + HEALTHCHECK in Dockerfile
- Add SIGTERM/SIGINT graceful shutdown handler (drains connections, closes DB)
- Run container as non-root sencho user in Dockerfile
QUALITY:
- Fix 4 silent empty catch{} blocks in EditorLayout (now show toast.error)
- Connect ErrorBoundary to root App in main.tsx
- Replace WebSocket.Server with named WebSocketServer import (ESM compat)
TESTING (new automated test suite):
- Install Vitest; 38 backend tests across 4 suites covering validation utilities,
health endpoint, auth middleware, login flows, SSRF protection, and path traversal
- Extract isValidStackName/isValidRemoteUrl/isPathWithinBase to utils/validation.ts
- Playwright E2E scaffolding: auth, stacks, nodes specs + shared login helper
- CI: run Vitest + ESLint on every PR
Remote-node alerts now appear in the local notification bell alongside
local ones, each tagged with the originating node name.
- backend: reorder WS upgrade handler so /ws/notifications?nodeId=<remote>
falls through to the existing proxy path instead of short-circuiting
- frontend/api.ts: add fetchForNode() helper for explicit node-targeted
requests without touching the localStorage active-node key
- frontend/EditorLayout: fetch notification history from all registered
nodes in parallel on mount and on node-list changes; open a per-remote
WebSocket connection for real-time push; route mark-read / delete / clear
actions back to the originating node; show node-name badge on remote alerts
When the gateway proxied a WebSocket upgrade for a remote node's interactive
terminal (host console or container exec), it injected the long-lived api_token
(scope: 'node_proxy') as the Bearer token. The remote's WS upgrade handler
correctly blocked this token via its isProxyToken guard (added in a prior
security hardening commit), causing a 403 → socket destroy → "connection error,
session ended" on the client.
Fix: introduce a POST /api/system/console-token endpoint that issues a
short-lived JWT (scope: 'console_session', 60 s TTL). Before forwarding an
interactive WS upgrade to a remote node, the gateway calls this endpoint using
the api_token, then substitutes the returned console_session token as the Bearer
header for the WS upgrade. The remote's isProxyToken guard (scope === 'node_proxy')
continues to block the long-lived api_token from spawning shells directly while
allowing gateway-delegated console sessions through.
Non-interactive WS paths (stack logs) continue to use the api_token unchanged.
The previous fix attempted to re-stream req.body via proxyReq.write() in the
proxyReq event handler. This raced against http-proxy's synchronous
req.pipe(proxyReq) call: when express.json() consumed the IncomingMessage stream,
Node.js's pipe() detected readableEnded and scheduled process.nextTick(proxyReq.end),
which fired before the proxyReq socket event. Any subsequent write() threw
"write after end", leaving the remote server stalled waiting for body bytes that
never arrived (the "Add Rule spins forever" bug).
Root fix: a conditional middleware now skips express.json() entirely for requests
targeting a remote node on non-auth, non-nodes API paths. The raw stream is left
intact, allowing http-proxy's req.pipe(proxyReq) to forward the body correctly.
express.json() fully drains the incoming request stream. http-proxy-middleware then
pipes an already-consumed empty stream to the remote Sencho instance, which holds the
connection open waiting for body bytes that never arrive (Content-Length says N but 0
bytes are forwarded). This caused all POST/PUT/PATCH requests to remote nodes — most
visibly 'Add Alert Rule' — to hang indefinitely with an infinite loading spinner.
Fix: in the proxyReq handler, write JSON.stringify(req.body) with correct Content-Type
and Content-Length headers before forwarding, restoring the full request body.
- StackAlertSheet: fetch agent status on open and show contextual banner
(amber warning when no channels configured, green when active, blue info
callout on remote nodes explaining remote-instance evaluation semantics)
- StackAlertSheet: surface actual server error message on addAlert failure
instead of generic string; add console logging for all failure paths
- SettingsModal: expose Notifications tab for remote nodes so agents can be
configured directly on any Sencho instance; section header shows node name
and Remote badge with tooltip when proxying to a remote node
- SettingsModal: re-fetch agents/settings when active node changes
- SettingsModal: add hidden DialogTitle/DialogDescription to satisfy Radix
UI accessibility requirements (eliminates console errors)
- backend POST /api/alerts: add Zod validation schema; rejects unknown
metric/operator values, negative thresholds, and missing fields with 400
- EditorLayout WS: upgrade reconnect from flat 5s retry to exponential
backoff (1s→2s→4s→8s→16s→30s max); onerror now logs the event; cleanup
only closes OPEN sockets — CONNECTING sockets are closed in onopen after
isMounted check, eliminating "closed before established" StrictMode error
- Reject node_proxy scoped JWT tokens with 403 on host-console and
container exec (/ws) upgrades; machine-to-machine credentials must
not open interactive shells
- Validate stackParam against path.resolve + startsWith(baseDir) to
prevent directory traversal on the PTY cwd (Rule 9 pattern)
- Strip JWT_SECRET, AUTH_PASSWORD, AUTH_PASSWORD_HASH, DATABASE_URL
from the environment passed to node-pty spawned shells
Install motion + animate-ui, overhaul design tokens with brand cyan accent,
and replace CSS keyframe animations in Dialog, Tabs, Switch, and Tooltip
with spring-physics and blur-fade transitions via animate-ui Radix primitives.
Eliminates the 5-second setInterval polling loop for notifications in
EditorLayout. A persistent /ws/notifications WebSocket connection is
opened on mount; the backend pushes each alert the instant dispatchAlert
fires, with 5s auto-reconnect on close.
Key changes:
- NotificationService: injectable broadcaster callback (setBroadcaster)
- DatabaseService.addNotificationHistory: returns full NotificationHistory
record (with id/is_read) instead of void
- index.ts: notificationSubscribers Set + /ws/notifications upgrade handler
(JWT-verified, placed before remote proxy path)
- EditorLayout: polling removed, WS connect/reconnect replaces it
Security:
- Strip auth credential keys (auth_username, auth_password_hash,
auth_jwt_secret) from GET /api/settings response
- Add allowlist guard to POST /api/settings — rejects unknown or
auth-namespace keys with a 400
Backend:
- Add PATCH /api/settings bulk endpoint with Zod schema validation
(type coercion, range checks, URL format) and atomic SQLite transaction
- Add system_state table — moves last_janitor_alert_timestamp out of
global_settings; adds getSystemState/setSystemState on DatabaseService
- Add metrics_retention_hours and log_retention_days configurable settings;
MonitorService reads both dynamically each evaluation cycle
- Add cleanupOldNotifications(days) to DatabaseService, called each cycle
Frontend:
- Replace single isLoading flag with per-operation states
(isSavingSystem, isSavingDeveloper, isSavingPassword, isSavingRegistry,
isSavingAgent/isTestingAgent per agent type)
- Add skeleton loader that blocks interaction until fetchSettings resolves
- Explicit key-picking in fetchSettings — auth keys cannot enter state
- Unsaved-changes amber dot on System Limits and Developer sidebar items
- Separate saveSystemSettings / saveDeveloperSettings — no cross-tab clobber
- Developer tab gains Data Retention section (metrics hours, log days)
- All settings saves use new PATCH /api/settings endpoint
Anso