🐧 SmartLinux

SmartLinux is a modern, high-performance desktop application for storage drive health diagnostics, telemetry inspection, and S.M.A.R.T. attribute monitoring across local drives (SATA, NVMe, USB) and remote SSH Homelab servers (Linux, macOS, BSD).

Designed with an on-demand philosophy, SmartLinux avoids unnecessary drive spin-ups and disk wear by only scanning when explicitly requested.

---

✨ Key Features

• ⚡ Pure On-Demand Diagnostics:
  Zero background spin-ups and zero unnecessary wear on standby HDDs. Storage devices and remote servers are listed immediately in an unscanned state; S.M.A.R.T. telemetry is read only when you click Scan Now. Selecting a drive (local or remote) only shows what is already known; once a drive has been read, the button becomes ↻ Refresh (also in the right-click menu) and the panel shows Last scanned with the time of that read.
• 🔌 Real-Time USB Hotplug Auto-Detection:
  Zero-overhead Linux kernel Netlink uevent listener automatically detects plugged or removed USB drives in real time without requiring manual refreshes.
• ⏏️ Safely Eject USB Storage:
  Safely unmounts and powers off external USB storage devices directly from the UI or right-click context menu.
• 🌐 Remote Homelab Monitoring via SSH:
  Connect seamlessly to remote servers (e.g., ZimaOS, Proxmox, TrueNAS, Raspberry Pi, Ubuntu Server, Debian, macOS) using Password or Private Key authentication with automatic sudo elevation.
• 🕰️ Legacy Device Support:
  Reads very old remote equipment (e.g. WD My Cloud NAS with smartctl 5.38, BusyBox, ssh-dss-only SSH). See Legacy devices.
• 🔌 USB Adapters That Hide SMART:
  Opt-in compatibility mode that briefly switches a UAS adapter to Bulk-Only to read the full SMART table, then switches it back. See USB adapters that hide SMART.
• 📄 Multi-Format Diagnostic Reports:
  Export comprehensive diagnostic reports with multi-drive selection and custom directory browsing:
  • 📝 Markdown (.md): Clean GitHub-Flavored Markdown tables.
  • 📕 PDF Document (.pdf): Styled high-DPI report with telemetry metric boxes and colored status badges.
  • 📘 Word Document (.doc): Formatted Microsoft Word / LibreOffice Writer report.
  • 📊 Excel Spreadsheet (.xls): Multi-sheet Excel workbook with per-attribute columns.
• 📈 On-Demand S.M.A.R.T. History Log:
  Persists health snapshots on every manual scan into a lightweight, non-database JSON Lines file (~/.local/share/smartlinux/history.jsonl). An interactive Historial tab computes delta comparisons against previous scans, instantly flagging worsening sectors or error counts in red.
• 🩺 Native Drive Self-Test (Short & Long):
  Trigger internal drive self-tests (smartctl -t short or -t long) directly from the diagnostic panel for both local and remote SSH drives, with asynchronous background polling and human-readable completion timing ("X hours/days ago").
• 🛡️ Hardened SSH Security & System Keyring:
  Replaces permissive auto-add policies with strict known_hosts cryptographic verification, protecting Homelab hosts from Man-in-the-Middle (MITM) attacks with interactive fingerprint approval modals. SSH passwords are saved securely in the system keyring (gnome-keyring / SecretService) with defensive 0600 / 0700 file permissions.
• 🤖 Standalone Ansible & ntfy Monitor:
  Includes scripts/ansible_smart_check.py, a standalone zero-dependency Python script designed for headless automation (e.g. Ansible / Semaphore). Evaluates exact ATA/NVMe thresholds and fires immediate push alerts to a self-hosted ntfy instance (NTFY_URL) on WARNING or FAILED status.
• 🔔 Opt-In Update Notice:
  Optionally tells you when a newer release is available. Off until you enable it; it only shows a version number and a link, and never downloads or installs anything. See Update notifications & privacy.
• 🎨 Modern High-Contrast Dark UI:
  Crisp typography (+15px readable scale), responsive telemetry cards (Temperature, Power-On Hours, Power Cycles), and full ATA / NVMe SMART attribute tables.
• 🪄 1-Click Dependency Auto-Configuration:
  If smartctl is missing or lacks non-root SUID permissions, SmartLinux presents a single-click auto-setup modal that installs packages and sets permissions via system authorization (pkexec).

---

🛠️ System Requirements

• Operating System: Linux (Ubuntu, Debian, Fedora, Arch, Linux Mint, Pop!_OS, etc.) or macOS.
• Python: Python 3.10 or higher.
• smartmontools: smartctl utility installed on the system.

Installing Dependencies

Ubuntu / Debian / Linux Mint / Pop!_OS:

sudo apt update
sudo apt install python3 python3-pip python3-venv smartmontools libxcb-cursor0

Fedora / RHEL / Nobara:

sudo dnf install python3 python3-pip smartmontools

Arch Linux / Manjaro:

sudo pacman -S python python-pip smartmontools

macOS (Homebrew):

brew install python smartmontools

---

🔒 Permission Setup (Local SUID)

To allow SmartLinux to read local raw disk health without prompting for root passwords on every startup, set the SUID bit on smartctl:

# Ubuntu / Debian / Fedora / Pop!_OS:
sudo chmod u+s /usr/sbin/smartctl

# Arch Linux / Omarchy / Manjaro:
sudo chmod u+s /usr/bin/smartctl

(Alternatively, let SmartLinux's built-in 1-click auto-setup modal handle this automatically upon initial launch).

---

🔔 Update notifications & privacy

SmartLinux can tell you when a newer release exists. This is off until you opt in:

• Opt-in: you are asked once at first launch. You can change it at any time from the Updates button in the status bar. Check for updates now always works on click, even with automatic checks off.
• What is sent: one anonymous HTTPS GET to https://api.github.com/repos/juniorbarchini-oss/smartlinux/releases/latest with the header User-Agent: SmartLinux/<version>. Like any web request, GitHub sees your IP address. No identifiers, no disk or server data.
• How often: at most once every 24 hours.
• What happens with the answer: nothing is downloaded or installed; SmartLinux only shows a version number and a link to the release.
• Where the setting lives: ~/.config/smartlinux/settings.json.
• No other telemetry.

---

🪟 Tiling Window Managers (Hyprland / Omarchy / Sway)

SmartLinux features a high-density diagnostic dashboard layout designed for a native window size of 1180x750. To prevent tiling window managers from auto-tiling or stretching the dashboard across ultra-wide monitors, add a floating rule:

Omarchy (~/.config/hypr/hyprland.lua):

o.window("smartlinux", { float = true, center = true, size = { 1180, 750 } })

Standard Hyprland (~/.config/hypr/hyprland.conf):

windowrulev2 = float, class:^(smartlinux)$
windowrulev2 = size 1180 750, class:^(smartlinux)$
windowrulev2 = center, class:^(smartlinux)$

---

🚀 Installation & Quick Start

Option 0: Distribution Packages

Ubuntu / Debian / Linux Mint / Pop!_OS:

Download smartlinux_<version>_all.deb from the GitHub release and install it:

sudo apt install ./smartlinux_<version>_all.deb

Arch Linux / Manjaro:

git clone https://github.com/juniorbarchini-oss/smartlinux.git
cd smartlinux/packaging/arch && makepkg -si

(AUR publishing is pending: AUR account registration is currently closed upstream.)

putty / putty-tools is optional and only needed for legacy SSH devices; see Legacy devices.

Option 1: Desktop Integration (1-Click Installer)

To install SmartLinux into your system's Application Menu with its official commercial icon:

git clone https://github.com/juniorbarchini-oss/smartlinux.git
cd smartlinux
chmod +x install.sh
./install.sh

Option 2: Run Directly (Development Mode)

git clone https://github.com/juniorbarchini-oss/smartlinux.git
cd smartlinux
chmod +x smartlinux.sh
./smartlinux.sh

Option 3: Compile Standalone Portable Package

To bundle SmartLinux into a standalone portable folder suitable for USB flash drives:

chmod +x build_portable.sh
./build_portable.sh

The compiled standalone binary will be generated in dist/SmartLinux/.

Releasing (maintainers)

1. Bump __version__ in smartlinux/__init__.py and pkgver in packaging/arch/PKGBUILD together (a test enforces that they match).
2. Tag vX.Y.Z and create the GitHub release.
3. Build the .deb with packaging/debian/build-deb.sh and attach it to the release.

---

🕰️ Legacy devices

SmartLinux also works with old NAS boxes and embedded devices (tested on a WD My Cloud EX4: Linux 3.2, BusyBox 1.20, OpenSSH 7.5, smartctl 5.38). Two independent problems are handled:

1. Old SSH servers (ssh-dss host keys)

paramiko ≥ 4 and OpenSSH ≥ 9.8 cannot negotiate ssh-dss. SmartLinux therefore has a second SSH engine based on PuTTY's plink, with the same behaviour for the rest of the app:

• Automatic: SmartLinux tries paramiko first; if the server fails with "no matching host key type", "Incompatible ssh peer" or another algorithm-negotiation error, it retries with plink.

• Manual: tick Legacy SSH (old devices) in the server dialog to use plink straight away.

• Install plink (optional, only for these devices; PuTTY 0.75+):

  System	Command
  Debian / Ubuntu	sudo apt install putty-tools
  Arch / Manjaro	sudo pacman -S putty
  Fedora	sudo dnf install putty
  macOS	brew install putty

• Security is not relaxed:

  • The password is handed to plink through -pwfile (a 0600 temp file that lives only as long as each plink process, preferably in $XDG_RUNTIME_DIR), never with -pw on the command line. -batch is always used.
  • Host keys are never trusted blindly. plink checks ~/.putty/sshhostkeys; for an unknown key SmartLinux shows the fingerprint and asks for confirmation. A confirmed fingerprint is pinned in ~/.config/smartlinux/legacy_hostkeys.json (0600) and enforced with -hostkey. A key that differs from the cached/pinned one aborts the connection with a MITM alert.

2. Old smartctl (no -j, no --scan)

• If smartctl -j answers UNRECOGNIZED OPTION, SmartLinux switches to text mode for that host and parses smartctl -a output (smartlinux/core/smart_text_parser.py): model, serial, firmware, capacity, health, hours, temperature, power cycles, the full ATA attribute table and the self-test log. It tolerates smartctl 5.x–7.x layouts, plus basic NVMe and SCSI/SAS reports.
• Drives are discovered without --scan or lsblk, from /proc/partitions and /sys/block/*/device/model.
• All -d device types (sat, usbjmicron, …) are tried in text mode too; smartctl -t reuses the one that worked.
• smartctl is also searched in /usr/local/modules/bin (WD firmware).
• Remote commands are BusyBox-ash compatible (no $((..)), no which, no lsblk).

Tests run offline against real captures in tests/fixtures/legacy_smartctl_5_38/.

---

🔌 USB adapters that hide SMART

Some USB adapters (e.g. the Seagate bridge in the DJI Fly Drive, USB ID 0bc2:a020) only answer basic SCSI commands while the kernel drives them with the fast UAS driver: SmartLinux then only gets "Healthy", with no temperature, power-on hours or attribute table. The drive is fine; the adapter just does not pass the SMART requests through. With the older Bulk-Only driver (usb-storage, kernel quirk VID:PID:u) the same drive returns the full ATA table.

When SmartLinux detects this (local USB disk, interface on uas, reading without a SMART table) it explains it in the Diagnostics tab and offers Read SMART in compatibility mode…. Nothing happens until you confirm a dialog that lists every step:

1. The drive's volumes are unmounted (udisksctl, as your user).
2. You are asked for administrator permission (pkexec, polkit action org.smartlinux.usbmode).
3. The packaged helper /usr/lib/smartlinux/usb-mode-helper adds only VID:PID:u to /sys/module/usb_storage/parameters/quirks (existing entries are kept, the previous value is saved in /run/smartlinux/) and re-attaches the USB port: ~20-30 s, the /dev/sdX name may change.
4. SMART is read (read-only, no self-tests).
5. The helper removes its entry and re-attaches the port again (always, also after an error), and the volumes are mounted again by UUID unless your desktop already did it.

SmartLinux refuses if a volume is in use (it shows which programs), if /, /boot, /home, swap or an LUKS/LVM volume lives on the disk, or if the disk is not on USB. Notes: the drive is slower while in compatibility mode; if SmartLinux is killed halfway the quirk stays until you reboot or replug the drive (or run pkexec /usr/lib/smartlinux/usb-mode-helper restore BUSID VID PID); the serial number reported differs between modes, so the reading shows up under its own entry in the History tab.

The helper and policy are installed by the .deb, the Arch package and install.sh (not when running from the source tree). Optional tools: udisks2, polkit/pkexec, lsof.

---

🗄️ RAID controllers (MegaRAID/PERC)

Behind a hardware RAID controller (LSI/Broadcom MegaRAID, Dell PERC) the operating system only sees virtual drives (/dev/sda, model e.g. "PERC H730 Mini"): they have no SMART of their own. The physical disks are only reachable as smartctl -d megaraid,N /dev/bus/H, which smartctl --scan lists.

The controller is a node of its own in the sidebar: "🟡 PERC H730 Mini (MegaRAID) — 5 physical, 4 virtual", with its physical disks under it and the virtual drives in a separate group. The dot is the worst of its disks and volumes (🟢 OK, 🟡 attention, 🔴 failure, ⚪ not scanned yet). Clicking the node opens the controller view (nothing is read just by selecting it):

• a header with the model, driver, firmware and battery, the overall status and the vendor tool status ("perccli64 007.1910 ✓", or what it adds and how to install it by hand when it is missing);
• ↻ Scan controller: reads everything in one go, in series (never in parallel, so neither the controller nor the SSH connection is flooded), with "Disk 3 of 5" progress and Cancel. First the RAID state with the vendor tool (once, if it is installed and permitted; without it, or without permission, a clear message and the disks are still read), then smartctl for every physical disk through the same per-disk read as a single scan, so the last-scanned time, history and cache are updated the same way. A disk that fails is marked with its error and the others are still read;
• a Volumes table (vendor tool data): name, RAID level, state, size, OS device and member disks; a RAID0 of one disk is marked "single-drive volume" (informational: it can be deliberate);
• one Physical disks table joining both sources per disk: slot/DID, model, SMART health, temperature, power-on hours, defects (grown defects / reallocated sectors), state on the controller and volume. Double-click a row to open the full disk view (which offers "← Back to controller"); each row's ↻ reads only that disk. Before the first scan the rows show "—" and a banner says so.

The controller view has three tabs, from the important to the detailed:

• Overview: one plain-language sentence on top that lists only the exceptions ("2 disks have reallocated sectors; 3 volumes have no redundancy.", or just "Everything is in order."), then the header, ↻ Scan controller, the Volumes and Physical disks tables described above.
• Maintenance: independent cards, each with a status dot, ONE status line and Show details (technical names as tooltips). Schedule: patrol read and consistency check (a consistency check that is not scheduled is shown as neutral information, not as an alarm) and the battery learn cycle. Battery: e.g. "Optimal · 98 % · protects the cache for ~32 min", with capacity, charge cycles, learn cycle and "replacement required". Controller: controller status and memory errors (uncorrectable or ECC errors are a warning). Operations: rebuild / initialization / consistency check in progress, with progress and time left only when the vendor tool reports them, or "No operation in progress". Everything comes from the vendor tool reads Scan controller already makes: no extra command.
• Trends: for each physical disk, how defects (grown defects / reallocated sectors), uncorrected errors and temperature changed in the last 7 and 30 days (▲ worse, ▼ better, = no change) next to the current value, from your local scan history. With fewer than two readings it says "Not enough history yet — scan again later". SmartLinux never predicts a time to failure.

The local history (~/.local/share/smartlinux/history.jsonl, one line per scan) also records, when available, grown_defects, uncorrected_errors and media_errors (vendor tool); older lines without these fields keep working.

SMART of the physical disks: smartctl -j -a -d megaraid,N /dev/bus/H (then -d sat+megaraid,N for SATA disks). SAS disks show health status, temperature, power-on time, start/stop cycles, grown defect list, corrected/uncorrected read/write/verify errors, non-medium errors and the self-test log. Health: a status other than OK is FAILED; grown defects or uncorrected errors are a WARNING (with the count); corrected and non-medium errors alone are normal on SAS. Empty slots are not listed.

The Server panel (select Local Drives or a server) shows the machine's identity (model, service tag with root, CPU, RAM, OS / Proxmox VE, kernel, uptime), its storage controllers, a short RAID controllers table (status and disk counts; double-click opens the controller), Export diagnostic bundle… and ↻ Scan all disks, which reads every physical disk of the machine in series (also on machines without a RAID controller). While a RAID controller has not been scanned yet, the panel says: "This server has a RAID controller. Open it and click Scan controller…". Scans are always started by you: nothing is read on a timer, on hotplug or just by selecting something.

---

🧰 Vendor tool (RAID)

smartctl reads each physical disk behind a MegaRAID/PERC controller, but not the RAID itself. When Dell perccli or Broadcom storcli is already installed, Scan controller uses it automatically (there is no setting to choose; the old vendor_tool_modes setting is ignored) to add:

• the controller: product, FW version, BIOS, driver, personality, controller status and battery (BBU: state, charge, temperature, replacement required, pack missing);
• every volume: RAID level, state (Optl / Dgrd / Pdgd / OfLn), size, the /dev/sdX it appears as, the physical disks it is built from and active operations (rebuild, init, CC). The sidebar shows the volume on each disk ("sdb · VMDATA1 · RAID0 · 1 disk · no redundancy");
• every physical disk: state on the controller, media / other error counts, predictive failure count, SMART alert. DID = the N of smartctl -d megaraid,N, so each row matches its sidebar disk.

Health: a volume Dgrd/Pdgd is a WARNING (redundancy lost), OfLn is FAILED; a disk Failed/Offln/UBad is FAILED, a predictive failure or a SMART alert is a WARNING, media errors are ATTENTION (other errors are informational); a battery that is not Optimal or needs replacement is a WARNING. The controller shows the worst of its parts. A missing component ("Cachevault doesn't exist") is not an error.

Policy:

• Read-only whitelist. Only these fixed commands run, with the controller number checked against ^[0-9]{1,2}$ and no shell (over SSH the validated tokens cannot contain shell syntax): -v, show ctrlcount J, /cN show J, /cN show all J, /cN/vall show all J, /cN/eall/sall show all J, /cN/bbu show all J, /cN/dall show J. Anything else (set, start, stop, add, del, import, clear, flash, secure, foreign, patrolread...) is refused before it runs. SmartLinux never changes the RAID configuration.
• Fixed executables: perccli64, perccli, storcli64, storcli, /opt/MegaRAID/perccli/perccli64, /opt/MegaRAID/storcli/storcli64.
• SmartLinux never installs it, locally or over SSH. If it is missing, the controller view says what it adds and where to get it (Dell: PERCCLI in Drivers & Downloads, category SAS RAID; Broadcom: StorCLI) and how to install it by hand (sudo dpkg -i perccli_*.deb / sudo rpm -ivh perccli-*.rpm).
• Manual only: it runs only when you click Scan controller. It needs root (or sudo for a remote user with a saved password); otherwise the view explains why and the disks are read with smartctl.

---

📦 Diagnostic bundle

When SmartLinux does not understand a machine (an HP, Lenovo or Dell server, a RAID card, a USB bridge), its owner can collect real output for the developer, with no remote access. In the Server panel (this computer or any server) click Export diagnostic bundle…:

1. The dialog lists exactly which commands will run. All of them are read-only, from a fixed list: /etc/os-release, uname -srm, lsblk -J -d -o NAME,PATH,MODEL,SIZE,TYPE,TRAN,ROTA,SERIAL,VENDOR,WWN, lspci -vmm -nn, smartctl --version, smartctl --scan (JSON and text), /sys/class/dmi/id/* (vendor, product, BIOS, board, chassis type), dmidecode -t system -t bios (only if already root), lscpu, MemTotal from /proc/meminfo, /sys/class/scsi_host/*/proc_name, modinfo (version and description only) of megaraid_sas, hpsa, smartpqi, aacraid, mpt3sas and nvme; for each disk the same SMART read the app uses (smartctl -j -a with its -d type, text -a on old smartctl, only smartctl -i for RAID virtual drives); and perccli / storcli only if already installed, through the Vendor tool whitelist.
2. Nothing runs until you click Collect and preview (it runs in the background, with Cancel). Each command has a 30 s timeout and a 2 MB output limit (truncated outputs are marked); a missing tool (e.g. on BusyBox) is recorded and the bundle goes on. Nothing is written on the inspected system and nothing is installed.
3. You see the sanitized files before anything is saved, then Save as… writes smartlinux-diag-YYYY-MM-DD-<model>.zip (no host name in the file name).

The zip contains README.txt, MANIFEST.json (command, exit code, duration, file, sha256, truncated), app_view.json (SmartLinux version, SSH engine, how the app classified each disk, its health and the controllers) and one file per command.

Privacy. SmartLinux sends nothing and opens no connection of its own: it only saves a local file. Before saving, personal values are replaced by consistent fake values (the same serial always gets the same fake serial, so files still match each other):

• serial numbers (smartctl, lsblk, dmidecode, vendor tool SN), WWN / SAS addresses / logical unit ids and NAA ids, UUIDs, service tags / SKU numbers and asset tags, MAC addresses, IPv4 / IPv6 addresses, host and server names (every occurrence), filesystem labels and RAID volume names, user names in paths (/home/<user>, /run/media/<user>); SCSI inquiry data (which embeds serials) is removed.
• Kept, because they are needed to fix bugs: models, vendors, firmware and software versions, smartctl -d types, sizes and SMART counters. Placeholders ("Not Specified", "0000…") and values shorter than 3 characters are not treated as personal.
• Defense in depth: after sanitizing, SmartLinux searches every output file (and the file name) for every original value it replaced, including serials hex-encoded. If any is left, the export stops with a message and no zip is written. The manifest only counts replacements per category; the original values are never stored.
• The bundle never contains passwords, keys or the contents of any disk.

---

📂 Project Architecture

smartlinux/
├── scripts/
│   └── ansible_smart_check.py  # Standalone headless SMART monitor with ntfy push alerts
├── smartlinux/
│   ├── core/
│   │   ├── models.py           # Data models (DiskInfo, SmartAttribute, ServerConfig)
│   │   ├── detector.py         # Local disk discovery and safe ejection
│   │   ├── history_manager.py  # On-demand JSON Lines health snapshot log (~/.local/share/)
│   │   ├── smart_parser.py     # ATA/SATA, NVMe and SAS/SCSI SMART JSON parser & threshold evaluator
│   │   ├── raid.py             # RAID controllers: -d megaraid,N disks, virtual drive detection
│   │   ├── system_info.py      # Read-only host identity and storage controllers (local / SSH)
│   │   ├── vendor_tools.py     # Optional perccli/storcli reads: whitelist, detection, RAID health
│   │   ├── controller.py       # RAID controller as a unit: its disks, aggregate status, status sentence
│   │   ├── maintenance.py      # Maintenance cards (schedule, battery, controller, operations)
│   │   ├── trends.py           # 7 / 30 day changes from the local history (no predictions)
│   │   ├── diagnostics.py      # Diagnostic bundle: fixed read-only commands, manifest, zip
│   │   ├── sanitizer.py        # Consistent fake values for personal data + residue check
│   │   ├── ssh_client.py       # Hardened paramiko client + automatic engine selection (AutoSSHClient)
│   │   ├── plink_client.py     # Legacy SSH engine on PuTTY plink (ssh-dss devices)
│   │   ├── remote_common.py    # Transport-independent smartctl / discovery logic (JSON + text mode)
│   │   ├── smart_text_parser.py # Plain-text smartctl parser for old smartctl 5.x
│   │   ├── watcher.py          # Netlink kernel uevent hotplug watcher
│   │   ├── usb_compat.py       # UAS adapters hiding SMART: detection, guards, compatibility-mode flow
│   │   ├── exporter.py         # Multi-format report exporter (.md, .pdf, .doc, .xls)
│   │   └── config_manager.py   # Secure server persistence with system keyring integration
│   ├── ui/
│   │   ├── theme.py            # Dark theme stylesheet (QSS)
│   │   ├── main_window.py      # Main window & background worker orchestration
│   │   ├── sidebar.py          # Device tree with health badges and context menus
│   │   ├── detail_panel.py     # Telemetry metric cards, self-test actions and attribute table
│   │   ├── server_panel.py     # Server identity card, storage / RAID controllers, Scan all disks
│   │   ├── controller_view.py  # RAID controller view: Scan controller, volumes, unified disk table
│   │   ├── serial_scan.py      # One-at-a-time scan worker (progress, cancel)
│   │   ├── diagnostic_dialog.py # Export diagnostic bundle: command list, sanitized preview, Save as
│   │   ├── history_tab.py      # Historical SMART attribute delta and degradation viewer
│   │   ├── server_dialog.py    # Remote SSH server configuration & host key approval modal
│   │   ├── export_dialog.py    # Multi-disk and multi-format export dialog
│   │   ├── compat_dialog.py    # Consent dialog for the USB compatibility mode
│   │   └── setup_dialog.py     # 1-click dependency & SUID installer modal
│   ├── assets/
│   │   ├── icon.svg            # Vector application icon
│   │   └── icon.png            # High-resolution raster application icon
│   ├── __init__.py
│   └── main.py                 # Application entry point
├── tests/
│   ├── test_security.py        # Automated tests for known_hosts, keyring, and permissions
│   ├── test_ansible_smart_check.py # Automated tests for ATA/NVMe thresholds and ntfy alerts
│   ├── test_smart_text_parser.py   # Text parser tests (real smartctl 5.38 captures in fixtures/)
│   ├── test_legacy_remote.py       # Simulated WD My Cloud EX4: discovery, text mode, self-test
│   ├── test_legacy_ssh_engine.py   # plink engine and paramiko→plink automatic selection
│   ├── test_usb_compat*.py         # USB compatibility mode on a fake sysfs (detection, guards, flow, UI)
│   ├── test_megaraid*.py           # Dell R630 / PERC H730 captures: discovery, SAS health, server info, UI
│   ├── test_vendor_tools.py        # perccli captures: whitelist, detection, RAID health
│   ├── test_controller_view.py     # Controller node / view, Scan controller, Scan all disks
│   ├── test_maintenance_trends.py  # Overview sentence, Maintenance cards, Trends, history fields
│   ├── test_sanitizer.py           # Synthetic personal values: replaced, consistent, residue check
│   ├── test_diagnostic*.py         # Bundle whitelist, R630 replay, zip / manifest, size cap, dialog
│   └── test_usb_mode_helper.py     # Root helper: argument checks, quirks editing, idempotence
├── smartlinux.sh               # One-click launcher script
├── install.sh                  # System desktop & icon installer
├── packaging/common/           # usb-mode-helper (root, via pkexec) + org.smartlinux.usbmode.policy
├── build_portable.sh           # Standalone binary compiler (PyInstaller)
├── smartlinux.desktop          # Linux desktop entry specification
├── requirements.txt            # Python dependencies (PySide6, paramiko, keyring)
└── README.md                   # Documentation

---

👥 Credits

• Concept, Design & Development: Humberto Barchini (HB) & Antigravity (AGY)
• License: Open Source under the MIT License.

---

🌟 Support & Donations

If you find SmartLinux useful for managing your storage hardware and Homelab servers, feel free to support the project: