New standalone Magisk / KernelSU-Next module that rejects TCP/UDP connections from selected UIDs to 127.0.0.1 / ::1 via iptables owner-match rules. Covers the VPN/proxy detection vector where an app probes well-known localhost ports (7890, 1080, etc.) via connect() — the observer gets ECONNREFUSED, indistinguishable from a real closed port. Rules live in a dedicated chain `vpnhide_out` / `vpnhide_out6` with a jump from OUTPUT, applied atomically via iptables-restore. Configured by /data/adb/vpnhide_ports/observers.txt (one UID per line, UID < 10000 filtered out for safety). service.sh re-applies at boot after netd finishes its own chain setup. uninstall.sh flushes on module removal. No C code, no per-kernel builds, no Rust FFI — just a shell script leveraging the iptables binary that ships with every Android ≥ 4. Verified on Pixel 8 Pro (Android 16) with iptables 1.8.11 legacy: observer UID gets REJECT, non-observer UIDs are unaffected.
2.9 KiB
portshide
Magisk / KernelSU-Next module that blocks selected apps from reaching
localhost ports. Used to hide locally-bound VPN / proxy daemons
(Clash, Sing-box, V2Ray, Amnezia, etc.) from apps that probe for them
via connect(127.0.0.1, PORT) / connect(::1, PORT).
How it works
A small shell script installs iptables / ip6tables rules inside a
dedicated chain vpnhide_out / vpnhide_out6:
iptables -A vpnhide_out -m owner --uid-owner <UID> -d 127.0.0.1 -p tcp -j REJECT --reject-with tcp-reset
iptables -A vpnhide_out -m owner --uid-owner <UID> -d 127.0.0.1 -p udp -j REJECT --reject-with icmp-port-unreachable
…for every UID listed in /data/adb/vpnhide_ports/observers.txt,
plus the same for ::1 via ip6tables. A jump from OUTPUT into the
dedicated chain is inserted exactly once (iptables -C guarded).
Observer apps receive ECONNREFUSED — indistinguishable from a real
closed port. netd's own chains are never touched; our chain lives
beside them and is idempotently re-applied on every config change.
Install
Pick vpnhide-ports.zip in the KernelSU-Next or Magisk manager and
install. Reboot not strictly required — rules apply on next boot via
service.sh, or immediately if you manage observers through the VPN
Hide app (it invokes vpnhide_ports_apply.sh via su).
Configuration
Managed by the VPN Hide app (Protection → Ports). Direct shell alternative:
# /data/adb/vpnhide_ports/observers.txt — one UID per line
10451
10422
Then:
su -c sh /data/adb/modules/vpnhide_ports/vpnhide_ports_apply.sh
Why just localhost, and why for selected apps only
- Banking / anti-censorship detection apps probe
127.0.0.1:7890,127.0.0.1:1080,127.0.0.1:8080etc. Blocking these globally would break the VPN client itself (it needs to bind / use them). Per-UID REJECT gives surgical control. - Blocking all ports on loopback for observers (rather than a port list) is safe for typical observer apps: banks, госуслуги, marketplaces, non-browser Yandex apps, VK — none legitimately use localhost.
- Browsers (Chromium-based) are the only category with some legitimate localhost use (dev tools, PWAs). Just don't add them as observers.
Caveats
- Rules are lost on reboot.
service.shrestores them early in boot, waiting fornetdto finish its setup first (checksbw_OUTPUTreadiness). - Some Android versions rebuild
OUTPUTon network state changes. Our rules in our own chain survive; only theOUTPUT -j vpnhide_outjump can be affected. Re-run apply script if needed; the VPN Hide app's Save action does this automatically. iptables-legacybackend expected (default on AOSP 16 as of this writing). nftables backend viaiptables-nftalso works — same syntax, same kernel effect.
Uninstall
Via root manager — uninstall.sh flushes and removes our chains.