A Cyberpunk 2077-style netrunner suite for Omarchy — fully functional, not a mockup.
Every panel is backed by real system data (psutil, ip -j, nmcli, ARP tables,
systemctl, journalctl, threaded ICMP/TCP scans).
Built for TITAN (Arasaka Security deck) — but runs on any Linux box with Python 3.11+.
███╗ ██╗███████╗████████╗██████╗ ██╗ ██╗███╗ ██╗███╗ ██╗███████╗██████╗
████╗ ██║██╔════╝╚══██╔══╝██╔══██╗██║ ██║████╗ ██║████╗ ██║██╔════╝██╔══██╗
██╔██╗ ██║█████╗ ██║ ██████╔╝██║ ██║██╔██╗ ██║██╔██╗ ██║█████╗ ██████╔╝
██║╚██╗██║██╔══╝ ██║ ██╔══██╗██║ ██║██║╚██╗██║██║╚██╗██║██╔══╝ ██╔══██╗
██║ ╚████║███████╗ ██║ ██║ ██║╚██████╔╝██║ ╚████║██║ ╚████║███████╗██║ ██║
╚═╝ ╚═══╝╚══════╝ ╚═╝ ╚═╝ ╚═╝ ╚═════╝ ╚═╝ ╚═══╝╚═╝ ╚═══╝╚══════╝╚═╝ ╚═╝ //OS
| Cyberpunk concept | Module | What it actually does |
|---|---|---|
| Cyberdeck status | dashboard |
per-core CPU, RAM, swap, disk, temps (acpitz/nvme), clock, load, uptime, GPU |
| Net access | netscan |
interfaces via ip -j, default route, DNS, nmcli, ARP table, threaded 254-host ping sweep |
| ICE | ice |
firewall probe (firewalld/ufw/nftables), all LISTEN sockets with owning processes, ICE strength rating on the Cyberpunk tier scale (White → Violet): firewall state and network-facing listeners; loopback-only sockets don't count |
| Daemons | daemons |
systemd roster (system + user), failed units, top processes |
| Data shards | shards |
journalctl entries parsed as shards with severity colors + a personal shard vault (~/.netrunner/shards/) |
| Breach Protocol | breach |
the actual 6×8 byte-buffer minigame: faithful row tables (2/3/3+4/3+4/5/5-byte sequences), cyan data / magenta malware / yellow maelstrom bytes, cursor movement, sequence execution, completion |
| Quickhacks | hack |
ping, traceroute, DNS decrypt (dig/getaddrinfo), threaded TCP port scan (42 common ports), HTTP inspect |
| Armory (damage) | armory |
executes real actions: kill processes (Downtime), sever TCP sockets via ss -K (Short Circuit), Wake-on-LAN, systemctl control, ssh exec, ping flood, reboot/suspend the deck — every action gated behind --yes/confirm and recorded in ~/.netrunner/actions.jsonl |
| Fixer (AI) | edgerunner / ask |
a nameable AI fixer backed by a local Halogen/OpenAI-compatible server; answers net/Linux/Omarchy questions with a live snapshot of your deck as context |
netrunner # interactive TUI (8 tabs, live refresh)
netrunner boot # animated cold-boot sequence (built from live data)
netrunner status # one-shot cyberdeck readout
netrunner scan # netscan incl. live ping sweep of your /24
netrunner scan --no-sweep # interfaces + ARP + nmcli only
netrunner ice # ICE analysis + listening sockets
netrunner daemons [--all] # systemd roster
netrunner shards [-n N] # journal as data shards
netrunner shard "Title" "body" # store a shard in your vault
netrunner hack ping 1.1.1.1 # quickhacks: ping | trace | dns | ports | http
netrunner hack ports host --ports 22,80,443
# ARMORY — the damage half (all actions need -y/--yes and land in the action log)
netrunner kill 1234 # DOWNTIME: signal a process (pid or name, -s KILL, -f)
netrunner cut 192.168.1.50 # SHORT CIRCUIT: sever TCP connections to a host (--port N)
netrunner wake aabb.ccdd.eeff # Wake-on-LAN magic packet (--via broadcast addr)
netrunner svc restart cups.service # daemon control (--user for user units)
netrunner ssh host uptime # remote exec (BatchMode, never hangs on password; -y to skip confirm)
netrunner flood host -c 64 # ping flood (ping -f w/ root, -i 0.1 fallback)
netrunner reboot -y / suspend -y # the deck itself
netrunner log # audit trail: every executed, refused or failed quickhack
netrunner breach # play Breach Protocol (WASD)
netrunner breach --demo --seed 42 # watch the deck auto-play
netrunner ask "why is my gateway unreachable?" # one-shot EDGERUNNER query
netrunner edgerunner # interactive chat with your fixer
netrunner config --operator V --corp ARASAKA --deck TITAN
netrunner config --fixer "Johnny Silverhand" # name your AI fixer
netrunner config --fixer-url http://192.168.1.20:8080 # point it at your server1–8switch tabs (Dashboard / Netscan / ICE / Daemons / Shards / Breach / Edgerunner / Armory)- arrows or
WASDdrive the Breach Protocol cursor nstarts a new protocol ·qjacks out- on the Edgerunner tab, type a question and hit
Enterto consult your fixer
NETRUNNER//OS is a per-user deck and refuses to launch under sudo
(NETRUNNER_ALLOW_ROOT=1 overrides). The few probes that need privileges —
firewall/ICE status via ufw/nft/firewalld — escalate themselves per
command with sudo -n, so the whole app never has to run as root. When a
backend is present but unreadable without privileges, ICE reports it as
UNKNOWN rather than falsely claiming the net is wide open.
install.sh adds an XDG desktop entry (~/.local/share/applications/NETRUNNER.desktop)
so the deck shows up in the Omarchy app launcher and opens tiled in a terminal
via xdg-terminal-exec --app-id=TUI.tile, with its own icon
(~/.local/share/icons/hicolor/scalable/apps/netrunner.svg). Without
xdg-terminal-exec it falls back to a standard terminal launcher.
edgerunner / ask talk to a local OpenAI-compatible inference server
(default http://localhost:8080; llama.cpp, LM Studio, Ollama, vLLM or Halogen all work)
— nothing leaves the LAN.
The model id is auto-detected from /v1/models (override with
--fixer-model). Each query carries a compact live snapshot of the deck
(interfaces, route, DNS, firewall state) so the fixer can reason about your
network; pass --no-context to skip it. Name the fixer whatever you like with
config --fixer; the endpoint can also be overridden per-run with
NETRUNNER_EDGERUNNER_URL / NETRUNNER_EDGERUNNER_NAME.
netrunner/
├── cli.py argparse front-end (every module reachable without the TUI)
├── ui.py Textual app: TabbedContent + auto-refresh panes + breach key routing
├── banner.py ASCII banner + boot sequence (generated from live interfaces/route)
├── palette.py cyberdeck color system + gauge/bar primitives + ICE tiers
├── config.py ~/.netrunner/config.toml (operator, corp, deck)
├── sysinfo.py psutil vitals, top processes, GPU probe
├── netinfo.py ip/nmcli/ARP parsing, ping sweep, traceroute, DNS, TCP scan, HTTP
├── ice.py firewall probe, listening sockets, ICE tier assessment
├── daemons.py systemctl parsing
├── shards.py journalctl → shards (handles byte-array + list MESSAGE encodings)
├── breach.py pure, testable game engine + CLI play/demo front-ends
└── edgerunner.py local AI fixer client (urllib → /v1/chat/completions)
Design rules followed:
- Rich 15-safe rendering — all dynamic text built via
Text.append(style=...), never markup strings, so log lines containing[UFW BLOCK]can't crash the renderer. - Engine/UI split —
breach.pyis pure logic (seeded, testable); the TUI and CLI are thin front-ends over it. - Graceless environments tolerated — every probe degrades: no dig → getaddrinfo, no firewalld → ufw → nftables, ufw needs root → "UNKNOWN, ICE presumed".
git clone https://github.com/vonhex/netrunner.git ~/Work/netrunner
~/Work/netrunner/install.shinstall.sh creates .venv (with uv if you have it, else python3 -m venv), links
~/.local/bin/netrunner, and adds the launcher and icon. Run it again after git pull to update.
Needs Python 3.11+.
.venv/bin/python -m unittest discover -s tests -p "test_*.py" # regression tests, pure logic
.venv/bin/python tests/smoke_test.py # headless TUI + breach engineThe regression tests need no root or network. The smoke test (app.run_test) mounts all
8 tabs, switches between them, drives the breach game with simulated arrow keys, and runs
5 seeded engine round-trips (generate → complete → render). Both run on every push and pull
request via GitHub Actions.
PolyForm Noncommercial License 1.0.0. Free to use, copy and modify for any non-commercial purpose: personal use, hobby projects, research, charities, schools and public bodies. Not for sale and not for commercial use: you may not sell NETRUNNER//OS, charge for it, or use it in or for a paid product or service. It's provided as is, with no warranty.
Only use the armory and scanning quickhacks on machines and networks you own or are allowed to test.