feat: standalone WireGuard from first install, fix networking stack

Standalone WireGuard (wg0:51820): - New archipelago-wg.service creates wg0 independent of NostrVPN - Keypair generated on first-boot, persisted on LUKS partition - vpn.create-peer uses wg genkey/pubkey (no nvpn dependency) - wg-address service depends on archipelago-wg, not nostr-vpn Networking fixes: - Remove nos.lol from default relays (requires PoW, events rejected) - Add Tor hidden service for private relay (port 7777) — NAT'd peers can reach relay over Tor for NostrVPN signaling - Fix Tor hostname sync race: wait loop before copying hostname files - Add tor-hostnames + wireguard dirs to LUKS partition setup - Include relay in hostname sync loops (setup-tor.sh + first-boot) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-08 20:27:38 +02:00
parent 5427d4ec5d
commit 2d1536f016
8 changed files with 112 additions and 129 deletions
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -1,25 +1,16 @@
 # CLAUDE.md — Archipelago (Archy)

-## Overview
-
 Archipelago is a **Bitcoin Node OS** — bootable, self-sovereign personal server. Flash to USB, install on hardware, manage via web UI.

 **Stack**: Rust backend + Vue 3 + TypeScript (strict) + Vite 7 + Tailwind + Pinia + Podman on Debian 12
 **Version**: 1.3.0 | **Target**: x86_64 and ARM64

---
-
 ## Beta Freeze (2026-03-18)

-**Phase 1: Feature Testing (internal) — WE ARE HERE**
-
-Feature set is LOCKED. Only: bug fixes, security hardening, ISO build fixes, UI polish, testing.
-No new features, no new apps, no new deps, no scope creep.
-
+Phase 1: Feature Testing (internal). Feature set is locked.
+Only: bug fixes, security hardening, ISO build fixes, UI polish, testing.
 Track: `docs/BETA-PROGRESS.md` | Checklist: `docs/BETA-RELEASE-CHECKLIST.md`

---
-
 ## Quick Reference

 ```bash
@@ -28,19 +19,6 @@ cd neode-ui && npm run build      # Build (outputs to web/dist/neode-ui/)
 ./scripts/deploy-to-target.sh --live  # Deploy to live server (.228)
 ```

-## Infrastructure
-
-| What | Where |
-|------|-------|
-| Dev server | `192.168.1.228` (SSH key: `~/.ssh/archipelago-deploy`) |
-| Secondary | `192.168.1.198` |
-| Git remote | `git.tx1138.com` (remote name: `tx1138`) |
-| App registry | `80.71.235.15:3000/archipelago/` (HTTP, insecure) |
-| CI runner | act_runner on .228, workflow: `.gitea/workflows/build-iso.yml` |
-| ISO builds | FileBrowser at `http://192.168.1.228:8083` → Builds/ |
-| SSH creds | Gitignored `scripts/deploy-config.sh` |
-| Web password | `password123` |
-
 ## Architecture

 ```
@@ -48,75 +26,22 @@ Debian 12
  ├── Podman (rootless, user archipelago)
  ├── Nginx (80/443 → backend, app proxies)
  ├── Rust Backend (core/) on 127.0.0.1:5678
-  │     ├── core/archipelago/     — Binary, RPC, auth, sessions
-  │     └── core/container/       — PodmanClient, manifests, health
  └── Vue.js UI (neode-ui/)
-        ├── src/api/rpc-client.ts — All backend communication
-        ├── src/stores/           — Pinia state
-        ├── src/views/            — Pages
-        └── src/style.css         — ALL styling (global classes only)
 ```

 **Data paths**: `/var/lib/archipelago/{app-id}/` (data), `/opt/archipelago/web-ui/` (frontend), `/usr/local/bin/archipelago` (binary)

 ## Critical Rules

-1. **Never build Rust on macOS** — deploy script handles cross-compilation via rsync + remote build
-2. **Always deploy after changes** — `./scripts/deploy-to-target.sh --live`
-3. **Frontend builds to `web/dist/neode-ui/`** — not `neode-ui/dist/`
-4. **Container images**: `scripts/image-versions.sh` is the single source of truth. All scripts use `$*_IMAGE` variables, never hardcoded registry paths.
-5. **Type-check before committing** — `cd neode-ui && npx vue-tsc -b --noEmit`
-
-## Frontend
-
- `<script setup lang="ts">` always — no Options API
- Global CSS in `style.css` — **never inline Tailwind**
- `.glass-button` for ALL buttons — `.gradient-button` is BANNED
- `.glass-card` for containers, `.path-option-card` for interactive cards
- `translateZ(0)` + `isolation: isolate` on glass elements (Chromium compositor fix)
- Pinia for state, typed RPC client, handle loading/error/empty states
-
-## Backend (Rust)
-
- No `unwrap()`/`expect()` — use `?` with `.context()`
- `tracing` for logging, never `println!` or log secrets
- Backend binds `127.0.0.1` only — nginx handles external access
- Validate all input before path construction — reject `..`, `/`, null bytes
- `tokio` runtime, timeouts on all external ops
-
-## Security (Post-Pentest)
-
- RBAC: explicit method allowlists, never prefix matching
- Session cookies: `SameSite=Lax; HttpOnly; Path=/`
- Rate-limit auth endpoints, rotate tokens after privilege escalation
- Validate redirect URLs with `isLocalRedirect()`, never `v-html` with user input
- Container security: drop ALL caps, add only required, `no-new-privileges`, memory limits, health checks
- See `.claude/rules/` for detailed crypto, API, container, and Bitcoin rules
-
-## ISO Build & CI
-
-CI builds on every push to `main` via git.tx1138.com Actions.
-
-```bash
-# Manual build on .228:
-ssh archipelago@192.168.1.228
-cd ~/archy/image-recipe
-sudo UNBUNDLED=1 DEV_SERVER=localhost BUILD_FROM_SOURCE=0 ./build-auto-installer-iso.sh
-```
-
-**Debugging fresh installs** — SSH in and check:
-```bash
-cat /var/log/archipelago-install.log              # Full installer output
-cat /var/log/archipelago-first-boot-diagnostics.log  # Service status, nginx, LUKS, etc.
-sudo archipelago-diagnostics                       # Re-run diagnostics anytime
-```
-
-**Kiosk**: X11 on VT7, console on VT1. `Ctrl+Alt+F1` for terminal, `Ctrl+Alt+F7` for kiosk.
-Toggle: `sudo archipelago-kiosk enable|disable|toggle`
+1. Do not build Rust on macOS — deploy script handles cross-compilation via rsync + remote build.
+2. Always deploy after changes — `./scripts/deploy-to-target.sh --live`
+3. Frontend builds to `web/dist/neode-ui/` — not `neode-ui/dist/`
+4. Container images: `scripts/image-versions.sh` is the single source of truth. All scripts use `$*_IMAGE` variables, not hardcoded registry paths.
+5. Type-check before committing — `cd neode-ui && npx vue-tsc -b --noEmit`

 ## App Integration Checklist

-When adding/fixing apps, check ALL of these:
+When adding/fixing apps, check all of these:
 - `core/archipelago/src/api/rpc/package/` — config, capabilities, deps
 - `neode-ui/src/views/marketplace/marketplaceData.ts` — marketplace entry
 - `image-recipe/configs/nginx-archipelago.conf` — proxy rules (HTTP + HTTPS)
@@ -128,3 +53,7 @@ When adding/fixing apps, check ALL of these:

 Commits: `type: description` (`feat:`, `fix:`, `docs:`, `refactor:`, `test:`, `chore:`, `perf:`)
 Push to: `git push tx1138 main`
+
+## Compact Instructions
+
+When compacting, preserve: list of modified files, test results, deploy target state, current branch, infrastructure IPs.
--- a/core/archipelago/src/api/rpc/vpn.rs
+++ b/core/archipelago/src/api/rpc/vpn.rs
@@ -41,8 +41,12 @@ impl RpcHandler {
        // Prefer onion (always works), fall back to direct IP
        let relay_url = relay_onion.clone().or(relay_direct.clone());

+        // Standalone WireGuard public key
+        let wg_pubkey = tokio::fs::read_to_string("/var/lib/archipelago/wireguard/public.key")
+            .await.ok().map(|s| s.trim().to_string());
+
        Ok(serde_json::json!({
-            "connected": status.connected,
+            "connected": status.connected || wg_ip.is_some(),
            "provider": status.provider,
            "interface": status.interface,
            "ip_address": status.ip_address,
@@ -53,6 +57,7 @@ impl RpcHandler {
            "configured": config.enabled,
            "configured_provider": format!("{:?}", config.provider).to_lowercase(),
            "wg_ip": wg_ip,
+            "wg_pubkey": wg_pubkey,
            "node_npub": node_npub,
            "relay_url": relay_url,
            "relay_onion": relay_onion,
@@ -373,44 +378,48 @@ impl RpcHandler {
        let params = params.unwrap_or(serde_json::json!({}));
        let name = params.get("name").and_then(|v| v.as_str()).unwrap_or("Mobile");

-        // Get server status for endpoint info
-        let status = vpn::get_status().await;
-        if !status.connected {
-            anyhow::bail!("NostrVPN is not running. Start VPN first.");
+        // Check that wg0 is up (standalone WireGuard)
+        let wg0_up = tokio::process::Command::new("ip")
+            .args(["link", "show", "wg0"])
+            .output().await
+            .map(|o| o.status.success())
+            .unwrap_or(false);
+        if !wg0_up {
+            anyhow::bail!("WireGuard (wg0) is not running. Wait for first-boot to complete.");
        }

-        // Generate a keypair for the new peer via nvpn keygen
-        let keygen = tokio::process::Command::new("nvpn")
-            .arg("keygen")
-            .output()
-            .await
-            .map_err(|e| anyhow::anyhow!("nvpn keygen failed: {}", e))?;
-
-        if !keygen.status.success() {
-            anyhow::bail!("nvpn keygen failed: {}", String::from_utf8_lossy(&keygen.stderr));
+        // Generate a keypair for the new peer using wg genkey/pubkey
+        let genkey = tokio::process::Command::new("wg")
+            .arg("genkey")
+            .output().await
+            .map_err(|e| anyhow::anyhow!("wg genkey failed: {}", e))?;
+        if !genkey.status.success() {
+            anyhow::bail!("wg genkey failed: {}", String::from_utf8_lossy(&genkey.stderr));
        }
+        let peer_private = String::from_utf8_lossy(&genkey.stdout).trim().to_string();

-        let keygen_output = String::from_utf8_lossy(&keygen.stdout);
-        let lines: Vec<&str> = keygen_output.lines().collect();
+        let mut pubkey_cmd = tokio::process::Command::new("wg");
+        pubkey_cmd.arg("pubkey");
+        pubkey_cmd.stdin(std::process::Stdio::piped());
+        pubkey_cmd.stdout(std::process::Stdio::piped());
+        let mut pubkey_child = pubkey_cmd.spawn()
+            .map_err(|e| anyhow::anyhow!("wg pubkey spawn failed: {}", e))?;
+        if let Some(ref mut stdin) = pubkey_child.stdin {
+            use tokio::io::AsyncWriteExt;
+            stdin.write_all(peer_private.as_bytes()).await?;
+            stdin.shutdown().await?;
+        }
+        let pubkey_out = pubkey_child.wait_with_output().await?;
+        let peer_public = String::from_utf8_lossy(&pubkey_out.stdout).trim().to_string();

-        // Parse private and public keys from keygen output (format: "private_key=<key>\npublic_key=<key>")
-        let parse_key = |line: &str| -> String {
-            if let Some(pos) = line.find('=') {
-                line[pos + 1..].trim().to_string()
-            } else {
-                line.trim().to_string()
-            }
-        };
-        let (peer_private, peer_public) = if lines.len() >= 2 {
-            (parse_key(lines[0]), parse_key(lines[1]))
+        // Read server's WireGuard public key (standalone WG key, then fall back to nvpn)
+        let server_pubkey = if let Ok(key) = tokio::fs::read_to_string("/var/lib/archipelago/wireguard/public.key").await {
+            key.trim().to_string()
        } else {
-            anyhow::bail!("Unexpected keygen output: {}", keygen_output);
+            vpn::read_nvpn_config_value("node", "public_key").await
+                .ok_or_else(|| anyhow::anyhow!("Cannot read server public key"))?
        };

-        // Get server's WireGuard public key from nvpn config
-        let server_pubkey = vpn::read_nvpn_config_value("node", "public_key").await
-            .ok_or_else(|| anyhow::anyhow!("Cannot read server public key from nvpn config"))?;
-
        // Detect host IP — prefer config, then nvpn, then system detection
        let host_ip = if self.config.host_ip != "127.0.0.1" {
            self.config.host_ip.clone()
--- a/core/archipelago/src/nostr_relays.rs
+++ b/core/archipelago/src/nostr_relays.rs
@@ -38,7 +38,7 @@ pub struct RelayStats {
 /// Default relays seeded on first use.
 const DEFAULT_RELAYS: &[&str] = &[
    "wss://relay.damus.io",
-    "wss://nos.lol",
+    "wss://relay.primal.net",
    "wss://relay.nostr.band",
    "wss://relay.snort.social",
    "wss://nostr.wine",
@@ -451,7 +451,7 @@ mod tests {
        let _ = load_relays(tmp.path()).await.unwrap();

        toggle_relay(tmp.path(), "wss://relay.damus.io", false).await.unwrap();
-        toggle_relay(tmp.path(), "wss://nos.lol", false).await.unwrap();
+        toggle_relay(tmp.path(), "wss://relay.primal.net", false).await.unwrap();

        let stats = get_stats(tmp.path()).await.unwrap();
        assert_eq!(stats.total_relays, DEFAULT_RELAYS.len());
--- a/image-recipe/build-auto-installer-iso.sh
+++ b/image-recipe/build-auto-installer-iso.sh
@@ -369,6 +369,7 @@ COPY archipelago-reconcile.timer /etc/systemd/system/archipelago-reconcile.timer
 COPY archipelago-tor-helper.service /etc/systemd/system/archipelago-tor-helper.service
 COPY archipelago-tor-helper.path /etc/systemd/system/archipelago-tor-helper.path
 COPY nostr-vpn.service /etc/systemd/system/nostr-vpn.service
+COPY archipelago-wg.service /etc/systemd/system/archipelago-wg.service
 COPY archipelago-wg-address.service /etc/systemd/system/archipelago-wg-address.service
 COPY nostr-relay.service /etc/systemd/system/nostr-relay.service
 COPY nostr-relay-config.toml /etc/archipelago/nostr-relay-config.toml
@@ -399,7 +400,8 @@ RUN systemctl enable NetworkManager || true && \
    systemctl enable archipelago-reconcile.timer || true && \
    systemctl enable archipelago-tor-helper.path || true && \
    systemctl enable nostr-relay || true
-# nostr-vpn and wg-address are enabled by first-boot after Nostr identity is generated
+# archipelago-wg + wg-address: enabled by first-boot after WG key is generated
+# nostr-vpn: enabled by first-boot after Nostr identity is generated
 # (env file doesn't exist until onboarding, so pre-enabling causes crash-loop)

 # Remove policy-rc.d so services can start on first boot
@@ -494,6 +496,10 @@ NGINXCONF
        echo "   Using nostr-vpn.service from configs/"
    fi

+    if [ -f "$SCRIPT_DIR/configs/archipelago-wg.service" ]; then
+        cp "$SCRIPT_DIR/configs/archipelago-wg.service" "$WORK_DIR/archipelago-wg.service"
+        echo "   Using archipelago-wg.service from configs/"
+    fi
    if [ -f "$SCRIPT_DIR/configs/archipelago-wg-address.service" ]; then
        cp "$SCRIPT_DIR/configs/archipelago-wg-address.service" "$WORK_DIR/archipelago-wg-address.service"
        echo "   Using archipelago-wg-address.service from configs/"
@@ -1357,11 +1363,14 @@ HiddenServicePort 4080 127.0.0.1:4080

 HiddenServiceDir $TOR_DIR/hidden_service_fedimint
 HiddenServicePort 8175 127.0.0.1:8175
+
+HiddenServiceDir $TOR_DIR/hidden_service_relay
+HiddenServicePort 7777 127.0.0.1:7777
 TORRC

 # Create hidden service dirs with correct ownership and permissions (700, not 750)
 # Tor refuses to start if permissions are too permissive
-for svc in archipelago bitcoin electrumx lnd btcpay mempool fedimint; do
+for svc in archipelago bitcoin electrumx lnd btcpay mempool fedimint relay; do
    mkdir -p "$TOR_DIR/hidden_service_$svc"
    chown debian-tor:debian-tor "$TOR_DIR/hidden_service_$svc"
    chmod 700 "$TOR_DIR/hidden_service_$svc"
@@ -1406,7 +1415,7 @@ done
 # Sync hostnames to backend-readable directory
 HOSTNAMES_DIR="/var/lib/archipelago/tor-hostnames"
 mkdir -p "$HOSTNAMES_DIR"
-for svc in archipelago bitcoin electrumx lnd btcpay mempool fedimint; do
+for svc in archipelago bitcoin electrumx lnd btcpay mempool fedimint relay; do
    if [ -f "$TOR_DIR/hidden_service_${svc}/hostname" ]; then
        cp "$TOR_DIR/hidden_service_${svc}/hostname" "$HOSTNAMES_DIR/$svc"
        echo "$(date): Synced hostname: $svc" >> "$LOG"
@@ -1965,7 +1974,7 @@ mkdir -p /mnt/target/var/lib/archipelago
 mount /dev/mapper/archipelago-data /mnt/target/var/lib/archipelago

 # Recreate directory structure on encrypted partition
-mkdir -p /mnt/target/var/lib/archipelago/{data,config,containers,secrets,tor,identities,lnd,nostr-relay,nostr-vpn}
+mkdir -p /mnt/target/var/lib/archipelago/{data,config,containers,secrets,tor,identities,lnd,nostr-relay,nostr-vpn,tor-hostnames,wireguard}
 mkdir -p /mnt/target/var/lib/archipelago/containers/storage
 mkdir -p /mnt/target/var/lib/archipelago/data/cloud/{Documents,Photos,Music,Videos,Downloads}
 # Copy relay config from rootfs (LUKS mount hides what the Dockerfile put there)
--- a/image-recipe/configs/archipelago-wg-address.service
+++ b/image-recipe/configs/archipelago-wg-address.service
@@ -1,7 +1,7 @@
 [Unit]
 Description=Assign WireGuard server address to wg0
-After=nostr-vpn.service
-Wants=nostr-vpn.service
+After=archipelago-wg.service
+Wants=archipelago-wg.service
 ConditionPathExists=/sys/class/net/wg0

 [Service]
--- a/image-recipe/configs/archipelago-wg.service
+++ b/image-recipe/configs/archipelago-wg.service
@@ -0,0 +1,14 @@
+[Unit]
+Description=Archipelago Standalone WireGuard (wg0)
+After=network-online.target
+Wants=network-online.target
+ConditionPathExists=/var/lib/archipelago/wireguard/private.key
+
+[Service]
+Type=oneshot
+RemainAfterExit=yes
+ExecStart=/usr/local/bin/archipelago-wg setup /var/lib/archipelago/wireguard/private.key
+ExecStop=/bin/bash -c 'ip link del wg0 2>/dev/null || true'
+
+[Install]
+WantedBy=multi-user.target
--- a/neode-ui/dev-dist/sw.js
+++ b/neode-ui/dev-dist/sw.js
@@ -82,7 +82,7 @@ define(['./workbox-21a80088'], (function (workbox) { 'use strict';
    "revision": "3ca0b8505b4bec776b69afdba2768812"
  }, {
    "url": "index.html",
-    "revision": "0.huo00jkc7v4"
+    "revision": "0.nnkdothias"
  }], {});
  workbox.cleanupOutdatedCaches();
  workbox.registerRoute(new workbox.NavigationRoute(workbox.createHandlerBoundToURL("index.html"), {
--- a/scripts/first-boot-containers.sh
+++ b/scripts/first-boot-containers.sh
@@ -76,7 +76,13 @@ fi
 # The backend reads from /var/lib/archipelago/tor-hostnames/{service} at startup
 TOR_HOSTNAMES="/var/lib/archipelago/tor-hostnames"
 mkdir -p "$TOR_HOSTNAMES"
-for svc in archipelago bitcoin lnd electrumx btcpay mempool fedimint; do
+# Wait for Tor to generate hostname files (setup-tor.sh may still be running)
+for attempt in $(seq 1 10); do
+  [ -f /var/lib/tor/hidden_service_archipelago/hostname ] && break
+  log "Waiting for Tor hostnames (attempt $attempt/10)..."
+  sleep 3
+done
+for svc in archipelago bitcoin lnd electrumx btcpay mempool fedimint relay; do
  for dir in /var/lib/tor/hidden_service_${svc}; do
    if [ -f "$dir/hostname" ]; then
      cp "$dir/hostname" "$TOR_HOSTNAMES/$svc" 2>/dev/null
@@ -86,6 +92,25 @@ done
 chown -R archipelago:archipelago "$TOR_HOSTNAMES" 2>/dev/null
 log "Tor hostnames populated: $(ls $TOR_HOSTNAMES 2>/dev/null | tr '\n' ' ')"

+# ── Standalone WireGuard: generate keypair and start wg0 ──────────────
+WG_DIR="/var/lib/archipelago/wireguard"
+mkdir -p "$WG_DIR"
+if [ ! -f "$WG_DIR/private.key" ]; then
+  wg genkey > "$WG_DIR/private.key" 2>/dev/null
+  chmod 600 "$WG_DIR/private.key"
+  wg pubkey < "$WG_DIR/private.key" > "$WG_DIR/public.key"
+  chown -R archipelago:archipelago "$WG_DIR"
+  log "WireGuard keypair generated"
+fi
+modprobe wireguard 2>/dev/null || true
+systemctl enable --now archipelago-wg 2>/dev/null || true
+systemctl enable --now archipelago-wg-address 2>/dev/null || true
+# Open firewall port for standalone WG
+if command -v ufw >/dev/null 2>&1 && ufw status | grep -q "Status: active"; then
+  ufw allow 51820/udp >/dev/null 2>&1 || true
+fi
+log "Standalone WireGuard (wg0:51820) started"
+
 # ── Private Nostr Relay: start for VPN signaling and general use ──────
 if command -v nostr-rs-relay >/dev/null 2>&1; then
  # Relay config is pre-installed by ISO at /var/lib/archipelago/nostr-relay/config.toml
@@ -153,12 +178,9 @@ NOSTR_PUBKEY=${NOSTR_PUBKEY}
 NVPNENV
    chmod 600 /var/lib/archipelago/nostr-vpn/env

-    # Load WireGuard kernel module
-    modprobe wireguard 2>/dev/null || true
-
-    # Start NostrVPN and WireGuard address services
+    # Start NostrVPN mesh service (standalone WG already started above)
+    systemctl reset-failed nostr-vpn 2>/dev/null || true
    systemctl enable --now nostr-vpn 2>/dev/null || true
-    systemctl enable --now archipelago-wg-address 2>/dev/null || true
    log "NostrVPN configured with node identity and started"
  else
    log "NostrVPN: no Nostr identity yet — will configure after onboarding"