- Python 98.1%
- Makefile 1.9%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| tests | ||
| ipwall.py | ||
| Makefile | ||
| README.md | ||
ipwall
A terminal UI for managing iptables / ip6tables — browse chains, add and
reorder rules, flip policies, watch counters, and undo mistakes, without
memorising netfilter syntax.
Single file, pure Python standard library. No build step, no dependencies.
┌─ Chains (4) ───────────┬─ INPUT · policy DROP · 6 rules ───────────────────────┐
│ INPUT DROP │ # PKTS BYTES RULE │
│ FORWARD DROP │ 1 1.2K 345K -i lo -j ACCEPT │
│ OUTPUT ACCEPT │ 2 890K 1.1G -m conntrack --ctstate RELATED,… │
│ SSH-GUARD ·2 │ 3 412 28K -p tcp -m tcp --dport 22 -j SSH-… │
│ │ 4 0 0 -p tcp -m multiport --dports 80,… │
└────────────────────────┴───────────────────────────────────────────────────────┘
Requirements
Python 3.9+ and the iptables binaries. Root, or a sudo that can reach them.
Install
sudo make install # -> /usr/local/bin/ipwall
sudo make install PREFIX=~/.local
make uninstall
make on its own lists every target. make check runs the smoke tests, which
use the in-memory demo ruleset and never open the live firewall. make demo
launches the UI against that same fake ruleset.
Or skip make entirely — it's one file with no dependencies:
chmod +x ipwall.py && sudo install -m 755 ipwall.py /usr/local/bin/ipwall
Usage
ipwall --demo # explore the UI against a fake ruleset — safe, no privileges
sudo ipwall # manage the live IPv4 ruleset
sudo ipwall -6 -t nat # start on the IPv6 nat table
sudo ipwall --watch 2 # refresh counters every 2 seconds
sudo ipwall --dry-run # change nothing; log what would have run (press L to see it)
Start with --demo. It is the real interface driven by an in-memory ruleset,
so every key does exactly what it will do for real.
Keys
↑ ↓ k j · PgUp/PgDn · g G |
move · page · first/last |
Tab · ← → · Enter |
switch pane · chains ⇄ rules |
t T · 1…5 · 6 |
next/prev table · jump to table · IPv4 ⇄ IPv6 |
/ · Esc · < > |
filter rules · clear filter · scroll rule text |
a · A · i |
add rule (guided form) · add raw spec · insert above cursor |
e · d · J K |
edit rule · delete rule · move rule down/up |
y · p · z |
yank rule · paste below · zero counters |
c · x |
hide counters · exact ⇄ human-readable counters |
P · n · X · R · F |
policy · new chain · delete chain · rename · flush |
u · Ctrl-R · r |
undo · redo · reload from kernel |
w · S · O · L |
write ruleset to disk · save snapshot · restore snapshot · command log |
? · q |
help · quit |
How it works
Rules are read with iptables -S, which prints them in the exact syntax you'd
type, so what the UI shows is what the kernel holds. Counters come from
iptables -L -n -v -x and are matched to rules by position.
Every mutating action snapshots the full ruleset (iptables-save) first, so
u restores the previous state exactly — including reorders, flushes, and
whole-ruleset restores. The snapshot stack holds the last 50 changes.
Adding a rule through the form (a) shows the exact command it will run as you
type it. If a field combination is invalid — ports without a protocol, a
missing target — it says so before anything is executed.
Safety notes
- Changes take effect immediately, as with
iptablesitself. They live in the kernel only; presswto write them where your distro reloads them at boot (/etc/iptables/rules.v4,/etc/sysconfig/iptables, …). - Setting the
INPUTorFORWARDpolicy toDROPcan lock you out of a remote session. The confirmation dialog warns you;uundoes it. --dry-runexecutes nothing and records every command it would have run.