On Linux, rm on a running binary works (process keeps its fd).
Then cp creates a new inode. Restart service after.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The running binary locks the file, causing 'Text file busy' on cp.
Stop the service, copy, then restart.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The full URL form was 404. The short form lets Gitea resolve from
its configured action sources (GitHub proxy). This worked for build #7.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The runner cwd is the workspace itself, so deleting it removes the
shell's cwd. cd to home first, then clean workspace before clone.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The actions/checkout@v4 action was 404 on git.tx1138.com causing
instant build failures. Use manual git clone for reliability with
host-mode runner.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
All container image references now pull from 80.71.235.15:3000/archipelago/
instead of Docker Hub and ghcr.io. image-versions.sh is the single source
of truth; all scripts use $*_IMAGE variables instead of hardcoded refs.
Files updated:
- scripts/image-versions.sh: central ARCHY_REGISTRY variable
- core/*/config.rs: registry whitelist includes app registry
- core/*/stacks.rs: Immich + Penpot stack images
- scripts/{first-boot,deploy-to-target,container-specs}.sh: use variables
- docker/*/Dockerfile: nginx base image from registry
- image-recipe/: ISO build, podman config, menu script
- scripts/{container-doctor,deploy-bitcoin-knots,fix-indeedhub,validate-app-manifest}.sh
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
All app images now pull from 80.71.235.15:3000/archipelago/
instead of Docker Hub / ghcr.io. Insecure registry config
baked into ISO for fresh installs.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Workflow builds both variants on push to main. Manual trigger
lets you choose bundled, unbundled, or both. ISOs auto-copied
to FileBrowser /Builds/ folder for easy download.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Podman was caching the rootfs Docker layers, meaning firmware packages
and sources.list changes were never picked up on rebuild. Force fresh
build every time since the rootfs tar is the real cache.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The sed commands to modify debian.sources DEB822 format were silently
failing — firmware packages never got installed. Replace the entire
sources config with traditional sources.list that explicitly includes
non-free-firmware component.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The shim (shimx64.efi.signed) was being installed as BOOTX64.EFI but it
tries to load a second-stage binary with a garbled name, causing
"Failed to open \EFI\BOOT\" errors on machines with Secure Boot disabled.
Fix: use grub-install --removable directly (unsigned GRUB as BOOTX64.EFI).
This works on all UEFI hardware. Users with Secure Boot must disable it.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
EFI boot fix:
- Shim needs grub.cfg in same directory to find the root partition
- Create minimal grub.cfg in /EFI/BOOT/ that chains to /boot/grub/grub.cfg
- Preserve unsigned GRUB as fallback for non-Secure-Boot systems
- Copy full chain to both /EFI/BOOT/ and /EFI/archipelago/ paths
- Log EFI directory contents for debugging
Firmware fix:
- DEB822 format sed was wrong — fix Components line replacement
- Add fallback sources.list entry to guarantee non-free-firmware repo
- Ensures firmware-realtek, intel-microcode actually get installed
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- dd zero the 1MB BIOS boot partition before formatting to prevent
kernel FAT-fs bread() errors during boot (sda1 had stale data)
- Add intel-microcode and amd64-microcode packages to suppress
TSC_DEADLINE and similar CPU firmware bug warnings on boot
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Dashboard.vue: move DashboardMobileNav outside <main> so position:fixed
isn't broken by will-change:transform on the perspective container
- Add container-specs.sh and reconcile-containers.sh utility scripts
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Use find instead of hardcoded filename for downloaded ISO detection
(wget may save with redirect filename or partial name)
- Fix color escape codes: use $'\033' syntax instead of '\033' for
reliable ANSI color rendering in installer output
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add firmware-realtek, firmware-iwlwifi, firmware-misc-nonfree to rootfs
(fixes missing r8169 NIC firmware on Dell and other common hardware)
- Enable non-free-firmware repo in rootfs Dockerfile
- Suppress os-prober GRUB warning (GRUB_DISABLE_OS_PROBER=true)
- Auto-eject USB boot media before reboot to prevent re-entering installer
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The ISO build was using Debian 13 (Trixie) as the live installer base
while the rootfs was built from Debian 12 (Bookworm). This caused:
- Debian 13 kernel/hostname/user in the live environment
- Squashfs errors on reboot from live-boot initramfs hooks
Fixes:
- Pin live ISO to Debian 12.10.0 (archive URL)
- Remove live-boot/live-config packages before initramfs regeneration
- Clean out any live-boot initramfs hooks/scripts from installed rootfs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Remove || exit 1 from health-cmd (redundant, breaks SSH heredoc)
- Use --health-cmd 'cmd' format (space, not equals) for proper quoting
- Simplify bitcoin health check to bitcoin-cli getnetworkinfo (no creds needed)
- Fix MariaDB health check nested quote issue
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Fix shell escaping in LND config sync block (single-quoted SSH context
doesn't need backslash-escaped dollars)
- deploy-tailscale.sh BUILD_SOURCE auto-detects Tailscale IP when LAN
unreachable (fixes "No binary on .228" error)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add --all as alias for --fleet
- Fleet deploy auto-detects Tailscale IP when LAN SSH fails
- Skip .198 gracefully when unreachable instead of failing
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Create scripts/smoke-test.sh for live server verification (7 checks)
- Document planned GitHub Actions CI/CD pipeline in docs/ci-cd-plan.md
- Integration tests deferred to future task (require test harness setup)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- F29-F32: Split 4 views (Marketplace 1293→505, Server 1132→486, Home 1059→394, AppDetails 1036→386)
- F20: Add aria-current="page" to Dashboard nav links
- F21: Add 150ms search debounce in Marketplace and Apps views
- F22: Reduce backdrop-filter blur to 8px on mobile for GPU performance
- F23: Track and clear WebSocket connect check interval in all paths
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- S10: Add warnings to silent health check failures in deploy scripts
- S11: Add trap cleanup for temp dirs in deploy and tailscale scripts
- S12: Quote 20+ critical unquoted variables across deploy scripts
- S13: Extract hardcoded IPs to deploy-config-defaults.sh
- S15: Add --memory=256m to UI container runs
- F16: Remove in-memory JWT, use cookie-only auth in filebrowser client
- F17: Add meta tag fallback for CSRF token in RPC client
- F19: Track and clear setTimeout in AppSession on unmount
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- I2: Add MemoryMax=4G, LimitNOFILE=65535, TasksMax=2048 to systemd service
- I3: Tor rotation keeps old service for 1h transition before cleanup
- R14: Replace .parse().unwrap() with .unwrap_or(localhost) in rate limiter
- R15: Replace 7 unwrap/expect in mesh protocol with proper error propagation
- R27: Add 10s timeouts to mesh Bitcoin RPC calls
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- F4: Fetch fresh server state after WebSocket reconnect
- F5: Guard message polling timer with auth check, stop on logout
- F6: Remove NIP-07 listener in appLauncher close()
- F7: Initialize audio player once to prevent listener stacking
- S3: Pin all container images to specific versions, create image-versions.sh
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- R6: Convert 6 std::fs calls in session.rs to tokio::fs async
- R7: Convert std::fs::read_to_string in docker_packages.rs to async
- R8: Convert 3 std::fs calls in port_allocator.rs to async, switch to tokio::sync::Mutex
- R9+R10+R11: Fix blocking I/O in node_message.rs and nostr_discovery.rs
- R12: Convert electrs_status.rs from sync TCP to async tokio::net with 5s timeouts
- R4+R5: Spawn periodic cleanup tasks for endpoint and login rate limiters
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- F1: Guard connectWebSocket against concurrent calls with isWsConnecting flag
- F2: Serialize mesh send operations with sendQueue to prevent fetchMessages races
- F3: Add global Vue error handler with toast notification
- S1: Replace sudo podman with podman across all scripts (rootless Podman)
- S2: Add health-cmd to all 40 container run commands in first-boot-containers.sh
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- R1: Add health RPC endpoint with crash recovery status, uptime, and version
- R2: Wrap all 5 Nostr client.connect() calls in 10s timeout
- R3: Make backup restore atomic with staging dir and rollback on failure
- I1: Add rate limiting, body size, and proxy timeouts to unauthenticated nginx endpoints
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
AppDetails.vue now checks Bitcoin sync progress for LND, ElectrumX,
BTCPay, and Mempool. Shows orange warning banner with sync progress
bar and block height when Bitcoin is still syncing. Users see clear
feedback instead of broken wallet connect pages.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- lnd.rs: check tor-hostnames readable copy, then /var/lib/tor/, then
legacy /var/lib/archipelago/tor/ with sudo fallback for each
- electrs_status.rs: same multi-path resolution for ElectrumX onion
- Both servers: created /var/lib/archipelago/tor-hostnames/ with readable
copies of onion addresses (avoids sudo on every API call)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace all `podman` CLI shell-outs with HTTP requests to the rootless
Podman API unix socket (/run/user/{UID}/podman/podman.sock).
Benefits:
- No process spawning overhead — direct HTTP over unix socket
- Structured JSON responses — no string parsing fragility
- Proper timeouts on all operations (5s connect, 30s default, 120s create)
- Health check method to verify socket availability
- Restart container as first-class operation
Still uses CLI for:
- Image pulls (streaming operation better suited to CLI)
- Container logs (raw text stream, not JSON)
The Podman socket is rootless (runs as archipelago user), local-only
(unix socket), and already behind our session auth in the backend.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Messages between federated nodes are now end-to-end encrypted:
- X25519 ECDH key agreement from existing ed25519 node identities
- HKDF-SHA256 key derivation with domain separation
- ChaCha20-Poly1305 authenticated encryption per message
- Random 12-byte nonce per message via OsRng (CSPRNG)
- Graceful fallback to plaintext if encryption fails
- Receiver auto-detects encrypted vs plaintext messages
The Tor transport was already encrypted (onion routing), this adds
application-layer E2E encryption so even a compromised receiving
backend can't read messages without the node's private key.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Messages persisted to disk (messages.json) — survive restarts
- Sent messages stored on backend via node-store-sent RPC
- Message deduplication (same pubkey + message within 30s)
- Max 200 messages in circular buffer
- Direction field (sent/received) for proper UI display
- Container doctor: prefer system Tor, remove archy-tor container
- Deploy torrc generator: read from tor-config/services.json,
web apps map port 80→local port for clean .onion URLs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Major changes:
- Full Tor hidden service management via systemd path unit pattern
(tor-helper.sh + archipelago-tor-helper.path/service) — respects
NoNewPrivileges=yes, no sudo needed from backend
- Container doctor: prefer system Tor over container, remove archy-tor
- Deploy script: fix torrc generation (read correct services.json path),
web apps map port 80→local port, enable both tor and tor@default
- Federation: server rename pushes name to peers via background sync
- Server name: fix root-owned file, optimistic store update
- Mesh: local echo for sent messages, sendingArch loading state
- Web5: Message button → Mesh redirect, node name lookup in messages
- PeerFiles: show DID not onion in header
- Connected Nodes: flex-1 instead of fixed max-h
- Toast notifications route to Mesh
- Deploy script: fix single-quote syntax in SSH block
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- ShareModal: strip leading / from filepath (was causing "absolute paths not allowed")
- Server.vue: Tor status in Local Network section now uses same source as header
- Both fixes needed for file sharing and Tor to work consistently
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Public Channel:
- "Archipelago" channel in Mesh — broadcasts to all federation peers over Tor
- Shows received messages from all peers with pubkey label
- Auto-polls every 15s for new messages
- Orange-branded channel icon with unread badge
- Send handler routes to Tor broadcast when arch channel is active
FileBrowser Auto-Login:
- All filebrowser-client methods now call ensureAuth() before requests
- Auto-authenticates with default credentials if not logged in
- Fixes "files don't work when FileBrowser hasn't been logged into"
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Marketplace picks up in-progress installs from WebSocket store even
if install was started before page was opened. Removed nested .git.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Federation page uses bg-web5.jpg background
- Invite code in full-screen modal with type label (Link/Peer)
- Join modal upgraded to full-screen with backdrop blur
- "Untrusted" renamed to "Blocked" in trust selector
- Your Nodes / Peers containers: max-h-[60vh] with inner scroll
- Server name from Settings shown on DID card + network map
- DID sync between Web5 and Federation on rotation
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Web5 loads node DID from backend on mount (authoritative, survives rotation)
- Federation rotation updates localStorage so Web5 picks up new DID
- Cloud peer names: peerDisplayName() "Node-XXXX" instead of raw DID
- Cloud hides onion addresses from peer cards
- Sync timeout increased to 180s with better error message
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- nodeName() shows friendly "Node-XXXX" instead of truncated DID
- nodeNameFromDid() for sync results lookup
- Map labels use node names
- Content filename validation: allow / for subdirectories (Music/song.mp3)
but still block .., \, null bytes, hidden files, absolute paths
- Increased filename max length to 512 for paths with subdirectories
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Onion address shows "Not visible to peers" for non-trusted nodes
- Resource usage and app list only shown for trusted nodes
- Deploy app already gated to trusted only
- Backend should also strip data in get-state (future: TASK)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- DID in glass-card top-right (desktop) / below title (mobile)
- Your Nodes + Peers in two-column grid (lg breakpoint)
- "Remove Dead Nodes" button for unreachable peers
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Title: "Federation & Peers"
- Your Node DID moved to top-right header row (desktop), below title (mobile)
- Copy button shows "Copied!" feedback for 2 seconds
- Removed "X federated nodes" from description, added count to section header
- Rotate button compact in header
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- "My Node Identity" card shows DID with copy button
- "Rotate DID" button opens modal with password confirmation
- Rotation generates new keypair, then auto-notifies all federation peers
- Shows success/failure count after notification
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- node.rotate-did: generates new Ed25519 keypair, signs rotation proof
with old key, overwrites identity files, requires password
- federation.notify-did-change: broadcasts rotation proof to all
trusted/observer peers over Tor
- federation.peer-did-changed: receiving side verifies rotation proof
against known pubkey before updating peer's DID
- Rate-limited: 3/600s for rotation, 5/60s for peer notification
- Signature verification uses ed25519_dalek (constant-time)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Part 1 — DID Persistence:
- Deploy script creates /var/lib/archipelago/identity/ directory
- First-boot script creates identity dir with proper ownership
- Identity load now logs pubkey to confirm persistence across restarts
Part 2 — Node Names:
- NodeStateSnapshot includes node_name field
- build_local_state() passes server name to sync responses
- update_node_state() stores peer's announced name on the FederatedNode
- Names propagate automatically during federation.sync-state
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Task 13: added archy-* prefix containers, mempool-api, UI containers
to SERVICE_NAMES filter — removes empty squares from My Apps grid
- Task 12: App Store card hover changed from white/10 to orange-500/5
with orange border glow (subtle, not severe)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Ensures LND Connect works through every deployment path:
- Nginx: CORS $http_origin on /lnd-connect-info (both HTTP+HTTPS)
- Nginx: no cookie gate (backend is 127.0.0.1-only)
- LND UI source: fetch with credentials: 'include'
- Deploy: rebuilds LND UI with --no-cache every deploy
- First-boot: --restart unless-stopped + memory limits on UI containers
- Backend: bound to 127.0.0.1:5678 in systemd service
Root cause was CORS: LND UI on :8081 fetching :80 is cross-origin.
Browser blocked reading the 200 response without CORS headers.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The LND UI runs on port 8081 (separate nginx container) but fetches
/lnd-connect-info from port 80. This is cross-origin, so browsers
block reading the response without CORS headers. Added dynamic
Access-Control-Allow-Origin from $http_origin.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- New Discover.vue (app store redesign)
- Fleet.vue dashboard for .228
- MeshMap.vue component
- Fixed Discover.vue type errors (unused var, type predicate)
- Various UI updates (Apps, Dashboard, Marketplace, Mesh, Web5)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Mandatory rules for all new code based on 33 pentest findings.
Covers: input validation, auth checks, SSRF prevention, session
management, CSP, nginx config, container security, RBAC.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
SameSite=Strict prevents cookies from being sent when iframe content
(like the LND UI at /app/lnd/) fetches endpoints on the parent origin
(/lnd-connect-info). Lax still protects against CSRF on POST requests
but allows same-site GET navigations and fetches from iframes.
This was the root cause of "Failed to fetch" on LND Connect.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Backend is bound to 127.0.0.1 — only nginx can reach it.
Nginx checks cookie_session presence. Adding backend auth broke
the LND UI iframe fetch because the session validation was too
strict for the cross-proxy cookie flow. The nginx layer is the
correct auth gate for this endpoint.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- New Discover.vue with hero banner, featured sovereignty stack apps,
principle cards, manifesto footer, and full app grid
- Featured apps (Bitcoin Knots, LND, BTCPay, Vaultwarden) with
expanded privacy/sovereignty descriptions
- Discover is first tab in categories bar on App Store pages
- Smart back navigation: detail pages return to Discover when navigated from there
- Category clicks from Discover navigate to Marketplace with category pre-selected
- Cypherpunk aesthetic: terminal tags, scanline overlays, gradient accents,
animated Bitcoin orange headings
- Global CSS classes: discover-hero, discover-terminal-tag, discover-featured-card,
discover-principle-card, discover-manifesto
- Route added: /dashboard/discover
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The previous blockchain.numblocks.subscribe call returned data in a
format the parser couldn't extract height from. headers.subscribe
returns {height: N, hex: "..."} which is properly parsed.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
ReceiveBitcoinModal was missing QR code generation that Web5.vue has.
Added canvas refs + qrcode rendering for both on-chain (bitcoin: URI)
and lightning (lightning: URI) receive flows. Matches Web5 pattern.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
LND was crash-looping because lnd.conf had 127.0.0.1:8332 (container
loopback, not reachable) and the old hardcoded password. Deploy script
now detects stale values and patches them to bitcoin-knots:8332 with
the current secrets file password. Fixes address generation failure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
TASK-17: Deploy script auto-tags successful clean deploys with next
alpha version number. Skips if commit already tagged or working tree
is dirty.
BUG-3: Updated IndeedHub submodule — removed dead nostrConfig with
hardcoded ws://localhost:7777 that caused WebSocket reconnection spam
in browser console. Relay detection via relay.ts (auto-detect /relay
proxy) is the active path.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Background task spawned on server startup: every 15 minutes, checks opt-in
status, builds anonymous health report (node ID hash, version, uptime,
CPU/RAM/disk %, container states, recent alerts), saves to disk, and POSTs
to TELEMETRY_COLLECTOR_URL env var if configured. Non-fatal on failure.
Fixed FiredAlert field references (kind not rule_type, timestamp not
fired_at) in both monitoring and analytics modules.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Tests expected router.push but panel mode (now default) uses panelAppId
store state instead. Updated assertions to check panelAppId. Fixed
BTCPay app ID from 'btcpay' to 'btcpay-server'. All 515 tests pass.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
