- Add parent_id to NodeUpdate schema in backend so PATCH /nodes/{id}
accepts it (was silently ignored before)
- Expose parent_id in update_node MCP tool schema so the MCP SDK
forwards it to the backend instead of stripping it
- Add regression tests in both backend and MCP layers
Exposes homelab topology to MCP-compatible AI clients (Claude Code, etc.)
over LAN via HTTP/SSE on port 8001.
- New mcp/ service: FastAPI + mcp SDK, SSE transport
- Auth: X-API-Key for AI clients, X-MCP-Service-Key for backend (Docker-internal)
- Resources: canvas, nodes, edges, scan/pending, scan/runs
- Tools: create/update/delete nodes+edges, trigger scan, approve/hide devices
- Backend deps.py: accepts JWT or MCP service key (no plain-text password)
- 40 tests (auth, resources, tools) across asyncio + trio
- docker-compose.yml: mcp service on port 8001
- README: MCP setup section with Claude Code/Desktop config examples
- Add animated toggle per edge in EdgeModal (cyan switch, "Flow Animation")
- SVG-native <animate> element for reliable cross-browser dot animation
- Persist animated field: backend model, schemas (EdgeBase/EdgeUpdate/EdgeSave), DB migration
- Include animated in App.tsx edgesToSave serialization so it survives save/reload
- Add animated: bool to EdgeData TypeScript type
On fresh installs the data/ dir may not exist, causing SQLite to create a
stub file without write permissions. mkdir(parents=True, exist_ok=True) runs
at import time before any DB operation.
Also add data/.gitignore to prevent homelab.db from ever being committed.
Backend:
- Add MAC OUI lookup table (QEMU 52:54:00, Proxmox bc:24:11, VMware, VBox, Hyper-V)
- suggest_node_type() now accepts mac param and uses OUI as a low-priority hint
Frontend:
- Detect VM type from MAC prefix on pending device cards
- Show colored badge (QEMU / PVE / VMware / VBox / Hyper-V) in orange with tooltip
datetime.UTC was introduced in Python 3.11. Users on 3.10 get an ImportError.
Also lower ruff/mypy target-version from py313 to py310 to match minimum supported version.
- Add AUTH_USERNAME, AUTH_PASSWORD_HASH (admin/admin) to root .env.example
- Remove duplicate backend/.env.example — root file is the single source of truth
- Fix fresh-install login failure caused by missing password hash
All settings (auth credentials, scanner ranges, status_checker interval)
now live in a single .env file via pydantic-settings. config.yml and
config.yml.example are deleted.
- Settings: add auth_username, auth_password_hash, scanner_ranges,
status_checker_interval; add load_overrides()/save_overrides() for
persisting runtime changes to data/scan_config.json
- auth.py: read credentials directly from settings
- scan.py: read ranges/interval from settings; write-back via save_overrides()
- scheduler.py: read interval directly from settings
- main.py: call settings.load_overrides() at startup
- docker-compose.yml: remove config.yml volume mount, add new env vars
- conftest.py: set settings fields directly instead of writing a temp config.yml
Login route reads config.yml for credentials but it is gitignored.
Session-scoped autouse fixture now writes a minimal config.yml to a
temp path and patches settings.config_path so test_scan.py auth
fixtures work in CI without a real config file.
- test_status.py: 8 tests covering WS auth rejection/acceptance, broadcast_status (dead connection removal, no response_time, no connections), broadcast_scan_update
- test_scheduler.py: 9 tests covering _load_interval variants, _run_status_checks DB update/last_seen/error handling, start/stop lifecycle
- scheduler.py: reinitialize AsyncIOScheduler on each start() to avoid stale event loop across test restarts
C2 - JWT token was stored in localStorage (XSS-accessible):
- Switch Zustand persist storage from localStorage to sessionStorage
- Token is now scoped to the current tab and cleared on browser close
H5 - docker-compose had unsafe SECRET_KEY fallback:
- Replace ${SECRET_KEY:-change_me_in_production} with :? syntax
- Docker Compose now aborts with a clear error if SECRET_KEY is unset
M1 - Login endpoint had timing leak allowing username enumeration:
- Always call verify_password() regardless of username match
- Use hmac.compare_digest() for constant-time username comparison
- Both checks run every time; attacker cannot distinguish wrong
username from wrong password via response timing
C3 - config.yml contains credentials, remove from git tracking:
- Add backend/config.yml to .gitignore
- git rm --cached to untrack it
- Add backend/config.yml.example with instructions
C1 - SECRET_KEY must come from .env, no unsafe default:
- Remove hardcoded "change_me_in_production" default from config.py
- App now fails to start if SECRET_KEY is not set (pydantic required field)
- Generate real random key in backend/.env (gitignored)
- Add backend/.env.example for new contributors
H1 - WebSocket /ws/status was unauthenticated:
- Backend: require ?token= query param, validate via decode_token(),
close with code 1008 (Policy Violation) if missing or invalid
- Frontend: append ?token=<jwt> to WebSocket URL
- NodeSave schema was missing custom_icon field — icon reset to default on reload
- handleSave was not including custom_icon in the payload
- CameraNode now uses Cctv icon (clearer than Camera)
- Added 'cctv' entry to ICON_REGISTRY for the icon picker
- New 'camera' node type with Camera icon (frontend + types)
- Scanner fingerprint: camera ports (554, 8554, 37777, 34567, 2020) now suggest 'camera' instead of 'iot'
- service_signatures.json: all camera/NVR entries updated to suggested_node_type=camera
- 'camera' added to priority list in suggest_node_type (above iot)
Banner-specific signatures (e.g. AdGuard on port 3000) were incorrectly
matching when nmap returned no banner, causing wrong suggested_node_type
(e.g. 'router' for a media server). Now a signature with banner_regex is
only matched when a banner is present AND matches the regex.
Also make _PORT_TYPE_HINTS always contribute to suggest_node_type found set
(not just as elif fallback) so well-known ports like 8006 still resolve to
proxmox even when a generic sig also matched.
- Add 65+ icons across 7 categories (Infrastructure, Media, Monitoring,
Storage, Security, Automation, Dev & Containers, Communications) covering
popular self-hosted apps: Home Assistant, Jellyfin, Plex, Grafana, Portainer,
Pi-hole, Vaultwarden, Gitea, Nextcloud, Node-RED, Frigate, etc.
- New nodeIcons.ts utility with ICON_REGISTRY, ICON_MAP and resolveNodeIcon()
- Inline icon picker in NodeModal: collapsible panel with search + grid grouped
by category; click to select, click again or Reset to revert to type default
- BaseNode uses resolveNodeIcon() so custom icon renders live on canvas
- Add custom_icon field to NodeData type, NodeBase/NodeUpdate schemas, Node ORM
model, and database.py idempotent ALTER TABLE migration
- Expand service_signatures.json from 35 → ~120 entries covering *arr apps,
smart-home (HA, MQTT, ESPHome), cameras (RTSP, Dahua, Tapo, Reolink),
network gear (MikroTik, UniFi, Pi-hole), auth (Authelia, Authentik, Vault),
monitoring (Grafana, InfluxDB, Loki, Uptime Kuma), containers (Portainer),
and many more home lab services
- Expand nmap port range to cover home lab ports (8989, 7878, 8123, 554, 1883, etc.)
and increase --host-timeout from 30s to 120s for reliable -sV detection
- Add _PORT_TYPE_HINTS fallback in suggest_node_type for ports without signatures
(cameras→iot, MQTT→iot, Proxmox→proxmox, MikroTik→router, etc.)
- Show actual port/protocol (e.g. TCP/9999) instead of "unknown_service" so
all open ports are visible even when not matched
- Update tests to reflect new unknown-port label format
Backend:
- asyncio.to_thread(_nmap_scan) — nmap no longer blocks the event loop
- Commit each discovered device immediately (previously one bulk commit at end)
- Update ScanRun.devices_found after each device so history panel shows live count
- broadcast_scan_update() pushes {type: scan_device_found} WS event per device
- Refactor broadcast_status to shared _broadcast() helper, adds type: "status" field
Frontend:
- useStatusPolling handles both WS message types (status / scan_device_found)
- canvasStore: scanEventTs + notifyScanDeviceFound() action
- PendingDevicesPanel auto-refreshes when WS scan event is received
- Add custom_color field to EdgeData type, Edge DB model and all schemas
- HomelableEdge applies custom_color as stroke override (before selected highlight)
- EDGE_DEFAULT_COLORS extracted to utils/edgeColors.ts (react-refresh compliant)
- EdgeModal: color picker row showing effective color (custom or type default)
with hex value, Reset button when custom color is active
- Auto-migration adds custom_color column to edges table
- Add resolveNodeColors() utility merging type defaults with per-node overrides
- Default colors per node type (cyan=isp/router/lxc/ap, green=switch/nas,
purple=server/vm, orange=proxmox, amber=iot, gray=generic)
- Remove glowColor prop from BaseNode — colors now come from node data
- ProxmoxGroupNode uses resolveNodeColors for group border/header/icon
- NodeModal: Appearance section with 3 color swatches (border, background, icon)
— click swatch to open native color picker; Reset to defaults button
- custom_colors persisted as JSON in DB (backend model + schemas + migration)
- 7 new unit tests for resolveNodeColors covering all node types + partial overrides
- Double-click any link to open Edit Link modal (type, label, VLAN ID, delete)
- Add updateEdge / deleteEdge actions to canvasStore
- Add container_mode field to proxmox nodes (backend model, schemas, migration)
- NodeModal shows container toggle for proxmox type (defaults ON)
- ProxmoxGroupNode renders as regular BaseNode when container_mode is OFF
- setProxmoxContainerMode store action handles structural changes atomically
(children parentId/extent, node dimensions)
- CanvasContainer filters edges between container proxmox and its children
- Canvas load respects container_mode when assigning parentId/extent
Root cause: handleAddNode/handleEdgeConfirm only updated Zustand store,
never creating records in the DB. canvasApi.save() only updated positions
of existing nodes, so nothing survived a page refresh.
Fix: save now sends the full canvas state (all nodes + edges + data) and
the backend does a full sync — upsert incoming, delete anything removed.
This means Save is the single source of truth: no need to call individual
create/update/delete APIs for every drag or edit.
- scanner.py: re-raise nmap exceptions so they reach ScanRun.error
(previously silently returned [] hiding the root cause)
- Sidebar: after triggering scan, auto-switch to History tab
- ScanHistoryPanel: auto-refresh every 3s while any run is 'running';
show error toast when a run transitions running→error; show spinner
on running runs; error message fully visible (no truncation)
- scripts/run_scan.py: standalone scan script to run with sudo for
nmap OS detection / SYN scans on macOS
FastAPI's redirect_slashes=True causes GET /api/v1/canvas to 307 redirect
to /api/v1/canvas/ — axios follows the redirect but drops the Authorization
header, resulting in 403. Fixed by declaring routes as empty string instead
of '/' so no redirect is issued. Same fix applied to nodes and edges routes.
Updated all tests to use paths without trailing slashes.
Frontend:
- Split nodeTypes/edgeTypes into separate .ts files (react-refresh)
- Remove setState-in-effect in NodeModal (key prop reset)
- Fix handleSave accessed before declaration (useRef pattern)
- Exclude src/components/ui/** from eslint (shadcn generated)
- Use defineConfig from vitest/config for test type support
Backend:
- ruff --fix: sort imports, datetime.UTC alias
- Raise line-length to 120, ignore E501 in tests
- Break long update_node/update_edge signatures
- pyproject.toml: per-file-ignores for tests
Pouzor