Though the same could be said of any software.
No doubt but it’s more comfort in a team than an individual.
Oh absolutely, and software that’s been around a long time (kernel, kde, gnome etc) have a proven track record but they were once all new projects and people trusted them.
That trust grew as they grew.
I implemented new features to address issues commonly reported on the forums:
Added & Hardened (Universal Orphan Triage & Zero-Residue Purge Engine)
- Dynamic 3-Tier Orphan Safety Classifier (
triage_orphan_packages):- Implemented an intelligent offline classifier analyzing unrequired packages (
pacman -Qtdq) via batch local ALPM metadata queries (LC_ALL=C pacman -Qi):
Tier 1 (Safe Leaves): Truly unrequired leaf packages (Optional For: None). Safe to purge automatically.
Tier 2 (Optional Dependencies): Packages actively utilized as optional dependencies by installed software. Surfaces exact reverse dependencies (e.g. Dolphin, GIMP, VLC) to prevent silent feature loss.
Tier 3 (Core & Toolchain Safety Guard): Regex blacklist preventing accidental deletion of kernel headers, firmware, GPU drivers, audio servers, fonts, and build toolchains (base-devel,rust,cargo,go,gcc,make, etc.).
- Implemented an intelligent offline classifier analyzing unrequired packages (
- Zero-Residue Cache Purge (Arch Wiki Maintenance Standard):
- Automatically executes
paccache -c "$pacman_cache_dir" --remove --uninstalled --keep 0immediately following orphan deinstallation (pacman -Rns) to eliminate dead.pkg.tar.zstarchives while preserving rollback history for installed packages (PACCACHE_INSTALLED_KEEP=2).
- Automatically executes
- Explicit Intent Protection (
pacman -D --asexplicit):- Integrated interactive action allowing users to mark useful build tools or dependencies as explicitly installed, permanently resolving recurring orphan alerts for user tools.
- Pre-Flight Gate 2 Orphan Advisory:
- Added passive informative advisory in
run_guarded_upgradewarning users about pending orphans and stale AUR compilation risks before initiating system upgrades (100% Upstream Harmonization: zero false alarm FAIL/WARN).
- Added passive informative advisory in
- Read-Only Audit & Table Reconstruction Fidelity:
- Added
check_orphan_packages()under System Health & Services inrun_health_checkand updatedreconstruct_tables_from_log()for 100% fidelity across report viewers.
- Added
- CLI & TUI Integration:
- Added
-o, --orphanscommand-line switch for standalone execution and interactive submenu under Option 6 (Safe Maintenance).
- Added
Added (Dynamic Mirrorlist Health & Atomic Ranking Engine)
- Real-Time Mirrorlist Health & Latency Probe (
check_mirrorlist_age):- Upgraded read-only audit check to evaluate active server redundancy and measure live TCP/TTFB round-trip latency to the primary Arch repository (
curl). - Differentiates lightning-fast regional mirrors (< 150ms), acceptable continental mirrors (150-400ms), and sub-optimal / cross-continental mirrors (> 400ms - raises
INFO ℹadvisory). - Flags dead primary mirrors (
WARN ⚠) while verifying general internet reachability. - Dynamic discovery across all active
/etc/pacman.d/*mirrorlist*topologies (Arch, EndeavourOS, CachyOS, Chaotic-AUR) with per-list server counts.
- Upgraded read-only audit check to evaluate active server redundancy and measure live TCP/TTFB round-trip latency to the primary Arch repository (
- Universal, Atomic Mirror Ranking Engine (
refresh_and_rank_mirrors):- Standalone engine supporting official
reflector, AURrate-mirrors, and distribution-specificeos-rankmirrors. - Atomic Staging Gate: Ranks to isolated temporary files (
mktemp), validates server count (>= 3), verifies live HTTP 200 reachability of the primary mirror, and atomically installs viainstall -m 644with automated.bakbackups and rollback safety.
- Standalone engine supporting official
- CLI & TUI Integration:
- Added
--mirrorscommand-line switch for standalone execution. - Added Option 3 (“Refresh & Rank Fastest Regional Mirrors”) and updated Option 4 (“Complete Maintenance”) in Main Menu Option 6 (Safe Maintenance & System Hygiene).
- Added
How do I get notified with new updates to the scripts?
Thank you as always for trying out my work! I haven’t implemented an update notification feature yet, and I’m not using GitHub Releases at the moment (I just overwrite the script file), so GitHub’s Watch feature won’t notify you. My main focus right now has been rewriting the script from being hardcoded to my personal configuration into a fully portable version. For now, you’ll have to check the repository manually, but I’ll look into adding a built-in update checker in the future.
Out of curiosity, do you use any AI agents on your machine (for example like me - for system troubleshooting)?
Yes. I do. Paid for Grok+ (20 coins/month) is the most capable for now.
Co-pilot has so many ethical restrictions and lagging behind the leaders. I have it via Copilot 365. yearly paid plan.
Gemini is trying to catch up. Freebie, but Google is spamming me to upgrade.
Meta Muse is like kid who wants to play with the major league Freebie, but he is dying to migrate my “life experience” to his db.
I do not have Claude, I know it is very capable but expensive.
The fluff that many humans used to call it noise, or harmless artifacts in the logs, Grok can call on the skilled agents in Linux for best course of action. I noticed when proposed solutions are not effective or further solutions cannot be proposed, the skilled agent would immediately recommend not chase cul-de-sac routes.
Exiting time, we are living in. The end is neigh.
Are you ready?
I haven’t tried Grok yet. If it allows you to generate an API key and use your subscription tokens in other agents, that’s a bargain.
You mentioned artifacts and hallucinations – this is exactly why this script exists. The more data the model has, the easier it is for it to decide on an action.
My current setup works like this: a basic admin skill (in Goose) analyzes the data gathered by sys-health, combined with knowledge of the local environment (down to the ISP name). This skill acts as the “hands” and uses the cheaper Gemini 3.8 Flash.
For complex tasks, there is an auditor skill based on Luna 5.6 (hands must provide certain tasks for audit). This is the “brain” – it only provides solutions or proposes new capabilities. The proposed solution then goes back to the “hands” for execution.
It’s all based on prepaid APIs, so it can get very expensive.
**Update: sys-health v2.32 verified and running smoothly on derivative distributions (Mabox Linux / Manjaro family)!** 
We recently conducted an architectural and blast-radius audit to ensure `sys-health` adheres to 100% dynamic universalism across the entire Arch ecosystem—not just on EndeavourOS.
Following real-world testing on Mabox Linux (Openbox/Manjaro base) with multiple kernels (`linux612`, `linux618`), we hardened several core components:
- **Multi-Distro Kernel & Bootloader Sync:** `sys-health` now maps distribution version-named artifacts (`vmlinuz-6.12-x86_64`, `initramfs-6.12-x86_64.img`) directly from mkinitcpio presets and module trees. Multi-kernel GRUB configurations pass cleanly with zero false-positive desync alerts.
- **Safe Maintenance & Deep Clean:** Desktop trash (FreeDesktop/Openbox), 17 browser caches, and systemd journal vacuuming work seamlessly, with strict safeguards protecting gaming shader caches (`mesa_shader_cache`, `GLCache`).
- **Pamac-Aware Safety Locks:** Maintenance and upgrade checks dynamically recognize active `pamac-daemon` processes to avoid database lock contention.
- **Guarded System Upgrade:** Pre-flight sanity checks (mount topologies, battery thresholds, DKMS headers) and post-flight bootloader/initramfs audits are fully compatible.
- **Distribution Mirror Guard:** Mirror ranking automatically routes to `pacman-mirrors` on Manjaro/Mabox to prevent pulling incompatible Arch upstream lists.
Big thanks to the community members testing on diverse setups (@Beiruty and github.com/musqz)! If you are running EndeavourOS, Arch, Manjaro, Mabox, or CachyOS, feel free to give it a testdrive:
https://github.com/Gbur79/Sys-Health
Test run
═══════════════════════════════════════════════════════════════════════════════════════╗
║ SYS HEALTH › Audit & Diagnostics ║
╚═══════════════════════════════════════════════════════════════════════════════════════╝
BOOT & CORE OS ✖
╭──────────────────────────────┬────────────────────────────────────────────────────────╮
│ Kernel & modules │ FAIL ✖ (linux-lts: missing kernel linux-lts: missing │
│ │ initramfs linux-cachyos: missing kernel linux-cachyos: │
│ │ missing initramfs linux-xanmod-edge-x64v3: missing │
│ │ kernel linux-xanmod-edge-x64v3: missing initramfs) - │
│ │ booted: 7.2.8-xanmod1-1-edge-x64v3 │
│ Bootloader sync │ INFO ℹ (systemd-boot: loader entries are not readable; │
│ │ run with sudo to audit boot entries) │
│ Initramfs │ FAIL ✖ (missing initramfs image) │
│ (linux-xanmod-edge-x64v3) │ │
│ EFI partition (/efi) │ PASS ✔ (mounted vfat, free: 759MB) │
│ Reboot pending │ PASS ✔ (running kernel is current) │
│ Previous session shutdown │ PASS ✔ (clean shutdown) │
╰──────────────────────────────┴────────────────────────────────────────────────────────╯
HARDWARE & DRIVERS ⚠
╭──────────────────────────────┬────────────────────────────────────────────────────────╮
│ GPU runtime │ PASS ✔ (AMD/ATI [amdgpu],AMD/ATI [amdgpu]) │
│ GPU errors & lockups │ WARN ⚠ (AMD GPU error in kernel log: Sep 27 14:26:13 │
│ │ eos kernel: amdgpu 0000:03:00.0: [drm] *ERROR* │
│ │ Unsupported screen format RA24 little-endian │
│ │ (0x34324152)) │
│ DKMS modules │ PASS ✔ │
│ CPU temperature │ PASS ✔ (49°C) │
│ SMART disk health │ PASS ✔ (4/4 OK) │
│ Power & Battery │ PASS ✔ (AC Desktop power) │
│ SSD/NVMe TRIM timer │ PASS ✔ (active) │
╰──────────────────────────────┴────────────────────────────────────────────────────────╯
SYSTEM HEALTH & SERVICES ⚠
╭──────────────────────────────┬────────────────────────────────────────────────────────╮
│ Root disk space │ PASS ✔ (Max: /mnt/eos_spare 29%) │
│ Systemd failed (system) │ PASS ✔ │
│ Systemd failed (user) │ PASS ✔ │
│ Magic SysRq keys │ PASS ✔ (full emergency control enabled) │
│ Pacman DB lock │ PASS ✔ │
│ Package file integrity │ WARN ⚠ (Corrupt: texlive-bin (missing: │
│ │ /usr/share/applications/xdvi.desktop)) │
│ .pacnew configuration files │ WARN ⚠ (2) │
│ Orphan packages │ INFO ℹ (1 unrequired - triage recommended) │
╰──────────────────────────────┴────────────────────────────────────────────────────────╯
NETWORK & UPDATES ✔
╭──────────────────────────────┬────────────────────────────────────────────────────────╮
│ Network link & Gateway │ PASS ✔ (eno1: 2.5Gb/s, gw: 192.168.1.254 0.6ms │
│ │ +IPv6:unreachable) │
│ System DNS │ PASS ✔ (15ms) │
│ Available updates │ PASS ✔ (none) │
│ Mirrorlist status │ PASS ✔ (Arch: 4d, 88ms, 10 srv │ Distro: 4d, 4 srv │ │
│ │ Chaotic: 4d, 10 srv) │
│ Arch News (Latest) │ PASS ✔ (Mkinitcpio >=42 requires manu…) │
│ Arch security audit │ PASS ✔ (0 actionable; 23 tracker backlog) │
╰──────────────────────────────┴────────────────────────────────────────────────────────╯
GAMING & STEAM READINESS ⚠
╭──────────────────────────────┬────────────────────────────────────────────────────────╮
│ Multilib repository │ PASS ✔ (Enabled in /etc/pacman.conf) │
│ Vulkan & 32-bit graphics │ PASS ✔ (RX 9070 XT (RADV GFX1201) - 64+32-bit RADV OK) │
│ Proton memory limits │ PASS ✔ (max_map_count: 1048576 - nofile: 2048) │
│ Kernel sync (fsync) │ PASS ✔ (futex_waitv syscall 449 verified) │
│ Kernel split-lock │ WARN ⚠ (Mitigation active and split-locks detected in │
│ │ dmesg) │
│ CPU governor & GameMode │ INFO ℹ (Governor: powersave - GameMode optional for │
│ │ stutter) │
│ Desktop session & GPU │ PASS ✔ (Session: wayland) │
│ Proton & Steam tools │ PASS ✔ (Steam OK - Valve Proton) │
╰──────────────────────────────┴────────────────────────────────────────────────────────╯
╔═══════════════════════════════════════════════════════════════════════════════════════╗
║ SYS HEALTH: ACTION REQUIRED ✖ ║
╚═══════════════════════════════════════════════════════════════════════════════════════╝Hey Beiruty! Please open an issue on GitHub so I can investigate this: https://github.com/Gbur79/Sys-Health/issues
It seems like you have an unusual multi-kernel setup (LTS + CachyOS + XanMod Edge x64v3 with systemd-boot), which is great for stress-testing “sys-health”.
Could you please open a GitHub issue and include the following?
- Your exact kernel setup / how the kernels were installed
- Bootloader and boot partition/EFI layout
- Output of “uname -r”
- Output of “sudo -v && echo “sudo OK””
- The relevant “/boot” and “/efi” contents
- Your “sys-health” version
- The full Kernel & Modules section from the report
Thanks for helping stress-test it!
I will do so soon.
Or howz about just another eos system cleaner or JAESC :>
Thanks for the invaluable contribution. Once you’re done reviewing the name, let me know if you find any of the actual features useful or broken.
Hi,
I have completed the project (currently addressing user-reported issues as they come up). The script has grown to 11,139 lines and includes extensive new functionality.
Could you please remove the flag from my topic? I need to edit the first post to add a proper description and a full list of its capabilities (https://github.com/Gbur79/Sys-Health/)