Web UI:
- Peer input Enter key now resolves ETH/@alias (like /peer command)
- ETH address stored and shown everywhere instead of raw fingerprint
- Call UI shows ETH address: "Calling 0x0021...", "In call with 0x9D70..."
- Server URL color: #444 → #666 (readable on dark background)
- Peer input placeholder: "ETH address, fingerprint, or @alias"
- peerEthAddr persisted in localStorage across sessions
Server:
- WS binary header: strip zero-padding from 64-char to 32-char fingerprint
- Call routing now works (was failing due to padded fingerprint lookup)
- startCall() resolves ETH/alias before sending CallSignal::Offer
- Audio bridge sends auth token to wzp-web as first WS message
- Deterministic room name: sorted fingerprint pair (both peers same room)
Docs updated:
- SERVER.md: WZP integration section (components, running, TLS, auth flow)
- USAGE.md: voice call usage for web and TUI
- LLM_HELP.md: call architecture, key files, environment vars
- LLM_BOT_DEV.md: note that bots cannot participate in calls
- TESTING_E2E.md: updated WZP prerequisites with correct flags
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Web client:
- After call goes "active", connects to WZP web bridge WS
- Mic capture: getUserMedia → ScriptProcessor → PCM int16 frames → WS
- Playback: WS → PCM int16 → Float32 → AudioContext.createBufferSource
- Room name derived from peer fingerprint (deterministic)
- Relay address fetched from /v1/wzp/relay-config
- Audio auto-starts on accept/answer, auto-stops on hangup/reject
- startAudio()/stopAudio() manage full lifecycle
TUI:
- /call shows "Audio: use web client for voice (TUI audio coming soon)"
- Signaling works, audio requires web client for now
This completes the last critical task — voice calls work end-to-end:
User A calls → signaling via featherChat WS → User B accepts →
both connect to WZP relay → audio flows
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- **bold**, *italic*, \`code\` rendered with ratatui styles
- # headers, > blockquotes, - bullet lists
- Multi-line messages split and indented per line
- Code spans: cyan bold, headers: white bold, quotes: gray italic
- No external dependency (custom md_to_spans parser)
- tui-markdown had ratatui version mismatch, built our own
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- ChatLine gains sender_fp field for tracking who sent each message
- App gains read_receipts_sent HashSet to avoid duplicate receipts
- After each draw(), visible received messages get a Read receipt sent
- Only fires once per message_id, skips system/self messages
- Sender sees blue ✓✓ (existing display logic already handles Read)
- All ChatLine literals across 6 files updated with sender_fp field
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Call state reload on restart:
- Loads Ringing/Active calls from sled into active_calls on startup
- Expires calls older than 24h automatically
TUI sender ETH cache prefill:
- prefill_eth_cache() resolves all known contacts on poll_loop start
- First message from known contacts now shows ETH address immediately
Server integration tests (10 new):
- push_to_client offline/online
- register_ws + connection cap (5 max)
- is_online + device_count
- kick_device + revoke_all_except
- deliver_or_queue offline/online
- call state lifecycle
- list_devices
155 tests passing (was 135)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
FC-P2-T6: /contacts shows online status (● online, ○ offline)
FC-P6-T6: Long messages word-wrap into multiple lines with aligned indent
FC-P6-T7: Tab completion for 33 slash commands (4 new tests)
FC-P8-T6: sendDocument accepts both JSON and multipart form data
OTPK: Auto-replenish on TUI startup when supply < 3 (generates 10 new)
135 tests passing (was 127)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
FC-P2-T4: TUI call state machine
- CallInfo struct + CallPhase enum (Calling/Ringing/Active)
- Header shows call indicator: yellow "Calling...", magenta "Incoming", green timer
- /call sets Calling, /accept sets Active, /reject+/hangup clears
- Incoming signals show contextual messages (Offer→prompt, Answer→connected, etc.)
FC-P2-T5: Missed call display in TUI
- WS Text frames parsed for missed_call + bot_message JSON
- Missed calls: "📞 Missed call from X at HH:MM" + terminal bell
- Bot messages rendered as @botname: text
FC-P8-T5: Inline keyboard buttons in web
- CSS styled keyboard buttons (.kbd-btn)
- Bot messages with reply_markup render clickable button rows
- Click sends callback_data back to bot as bot_message
- Works in both WS text handler and handleIncomingMessage fallback
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Web client:
- Call bar between header and messages (hidden when idle)
- Call button appears when peer is set (not group)
- Incoming call: pulsing notification + Accept/Reject buttons
- Call states: idle → calling → ringing → active
- /call, /accept, /reject, /hangup slash commands
- CallSignal sent via WS binary frames (same as messages)
- handleCallSignal processes Offer/Answer/Hangup/Reject/Ringing/Busy
- Vibration on incoming call (mobile)
- create_call_signal WASM import wired up
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Per-bot numeric IDs (privacy feature) broke sendMessage because the
reverse lookup couldn't find the fingerprint from the per-bot hash.
Fix: store numid:<numeric_id> → fingerprint in tokens tree when
generating updates. resolve_chat_id checks this mapping first.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Server:
- --bots-config <path> loads JSON array of system bots on startup
- Bots auto-created if missing, aliases restored on every start
- Bot list stored in DB for welcome screen (system:bot_list key)
- GET /v1/bot/list returns system bots (public, no auth)
Welcome screen:
- Web + TUI show available bots on first login
- "Available bots: @helpbot — featherChat help, @codebot — Coding..."
- Clickable in web (via address detection)
Config: bots.example.json with 10 suggested bots
Usage: warzone-server --enable-bots --bots-config bots.json
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Privacy: from.id is now Hash(bot_token + user_fp) → different bots see
different numeric IDs for the same user. Prevents cross-bot user correlation.
Removed id_str (raw hex fingerprint) from all bot API responses.
Updated LLM_BOT_DEV.md and LLM_HELP.md.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Built-in BotFather (Rust, server-side):
- Intercepts messages to @botfather in deliver_or_queue
- Commands: /newbot <name>, /mybots, /deletebot <name>, /token <name>
- Creates bot with fingerprint, token, alias, tracks ownership
- Replies via push_to_client or queue (works offline)
- Only active when --enable-bots is set
Standalone BotFather (Python):
- tools/botfather.py: uses bot API (getUpdates/sendMessage)
- Delegates core ops to built-in handler
- Extensible for additional features
- Reads token from BOTFATHER_TOKEN env or .botfather_token file
Flow: User messages @botfather → "/newbot MyBot" → gets token back
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
TUI:
- Header shows peer ETH address (resolved on /peer set)
- Own messages show ETH format
- Resolve display shows full formatted fingerprint (xxxx:xxxx:...)
- peer_eth field stored on App for header display
Web:
- Pasting 0x address in peer input box now resolves via /v1/resolve/
- Send path resolves 0x/@ before encrypting
- Click messages area → focuses text input
- Own messages show ETH format
Version: 0.0.23 → 0.0.24, SW cache wz-v4 → wz-v5
Build script: --local, --local-ship, --local-clean commands
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Version:
- Workspace + protocol: 0.0.21 → 0.0.22
- Web client VERSION: 0.0.17 → 0.0.22
- Service worker cache: wz-v2 → wz-v3
ETH identity:
- Added WasmIdentity::eth_address() export (derives from seed via secp256k1)
- Web client sends eth_address during key registration
- Identity display shows ETH address first, then fingerprint
- No more server-side resolve needed — computed client-side
CLAUDE.md:
- Added MANDATORY version bump rule (4 places to update)
- Must bump on every functional change, never skip SW cache
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Server-to-server communication via WebSocket at /v1/federation/ws
- Auth as first WS frame (shared secret), presence + forwards over same connection
- Auto-reconnect every 3s on disconnect, instant presence push on connect
- Replaces HTTP REST polling (no more 5s intervals, lower latency)
- Removed dead HMAC helpers (auth is now direct secret comparison over WS)
- Simplified ARCHITECTURE.md mermaid diagrams for Gitea rendering
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replaced workspace dep inheritance with explicit versions in
warzone-protocol/Cargo.toml. The crate now builds both as a
workspace member AND standalone.
WZP can now import warzone-protocol directly:
warzone-protocol = { path = "../featherChat/warzone/crates/warzone-protocol" }
This means WZP can delete its mirrored identity/crypto code and use:
- warzone_protocol::identity::{Seed, IdentityKeyPair, PublicIdentity}
- warzone_protocol::crypto::{hkdf_derive, aead_encrypt, aead_decrypt}
- warzone_protocol::ethereum::{derive_eth_identity, EthAddress}
- warzone_protocol::message::{WireMessage, CallSignalType}
- warzone_protocol::types::Fingerprint
Single source of truth for identity derivation — no more HKDF mismatches.
28/28 tests pass. Zero warnings.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
/file <path> now works in group mode (#group):
- Sends file header + chunks to each group member
- Same fan-out approach as group text messages
- Each member receives and reassembles independently
- Progress shown: "Sending 'file.pdf' to group #ops..."
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Storage:
- contacts sled tree: auto-tracked on send/receive
- fingerprint, alias, first_seen, last_seen, message_count
- history sled tree: all messages stored locally
- key: hist:<peer_fp>:<timestamp>:<uuid> for ordered scan
- sender, text, is_self, timestamp
TUI commands:
- /contacts or /c — list all contacts (sorted by most recent)
Shows alias, fingerprint, message count
- /history or /h — show last 50 messages with current peer
- /h <fingerprint> — show history with specific peer
Auto-tracking:
- On send: touch_contact + store_message (is_self=true)
- On receive: touch_contact + store_message (is_self=false)
- Both KeyExchange and Message variants tracked
Backup: contacts + history included in export_all (encrypted backup).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
TUI fixes:
- /r and /reply now work: tracks last_dm_peer from received messages
- /r switches peer to last DM sender, then type normally
- /p @alias works as shortcut for /peer @alias
- /eth shows Ethereum address in TUI
- /unalias removes your alias
Web fixes:
- /p @alias and /peer @alias resolve and set peer
- /r and /reply work (switch to last DM sender)
- /unalias removes alias
- /admin-unalias <alias> <password> for admin removal
- File download now shows as clickable link (not auto-download)
Server:
- POST /v1/alias/unregister — remove own alias
- POST /v1/alias/admin-remove — admin removes any alias
- WARZONE_ADMIN_PASSWORD env var (default: "admin")
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Files now appear as a styled clickable link in chat:
📎 filename.pdf (1.6 KB) from sender
Click to download. No auto-save dialog.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Aliases:
- /unalias — remove your own alias
- /admin-unalias <alias> <password> — admin removes any alias
- Admin password via WARZONE_ADMIN_PASSWORD env var (default: "admin")
- POST /v1/alias/unregister + POST /v1/alias/admin-remove
Reply:
- /r or /reply — switches peer to whoever last DM'd you
- lastDmPeer tracked on both web and TUI
- Then type normally to reply
Web:
- Version bumped to 0.0.15 (was stuck at 0.0.10)
- WASM rebuilt with latest protocol
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Server:
- Register stores per-device bundles: device:<fp>:<device_id>
- GET /v1/keys/:fp/devices lists all registered devices
- WS already pushes to ALL connected devices per fingerprint
- DB queue: first device to poll gets messages (acceptable for Phase 2)
Multi-device flow:
- Same seed on two devices → same fingerprint
- Both register with different device_ids
- Both connect via WS → both receive messages in real-time
- Each device maintains its own ratchet sessions
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Server:
- POST /groups/:name/leave — remove self from group
- POST /groups/:name/kick — creator can kick members
- GET /groups/:name/members — list with aliases + creator badge
CLI TUI:
- /gleave — leave current group
- /gkick <fp_or_alias> — kick (creator only)
- /gmembers — show member list with aliases and ★ for creator
Web client:
- Same commands: /gleave, /gkick, /gmembers
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Server:
- DedupTracker in AppState: bounded HashSet (10,000 IDs, FIFO eviction)
- send_message: extracts message ID from bincode, drops duplicates
- WS handler: dedup on both binary and JSON message frames
- extract_message_id() parses all WireMessage variants
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Protocol:
- WireMessage::Receipt { sender_fingerprint, message_id, receipt_type }
- ReceiptType enum: Delivered, Read
- id field added to KeyExchange and Message variants
- Receipts are plaintext (not encrypted) — contain only ID + type
Web client:
- Auto-sends Delivered receipt on successful decrypt
- Tracks sent message IDs with receipt status
- Displays: ✓ (sent, gray), ✓✓ (delivered, white), ✓✓ (read, blue)
- Receipt indicators update live via DOM reference
CLI TUI:
- Auto-sends Delivered receipt back to sender on decrypt
- Tracks receipt status per message ID
- Displays receipt indicators after sent messages
WASM:
- create_receipt() function for web client
- encrypt_with_id/encrypt_key_exchange_with_id for tracking
- decrypt_wire_message handles Receipt variant
17/17 protocol tests pass. Zero warnings.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
poll_loop now:
1. Tries WebSocket connection to /v1/ws/<fingerprint>
2. On success: receives messages in real-time (instant push)
3. On disconnect: reconnects after 3 seconds
4. On WS failure: falls back to HTTP polling every 2 seconds
Refactored message processing into shared functions:
- process_incoming() handles raw bytes
- process_wire_message() handles deserialized WireMessage
- Used by both WS and HTTP paths
Both CLI TUI and web client now use WebSocket:
- No more HTTP polling spam in server logs
- Messages arrive instantly on both clients
- HTTP poll kept as fallback for scripts/mules
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Server:
- WS endpoint: /v1/ws/:fingerprint
- Connection registry in AppState (fingerprint → WS senders)
- On connect: flushes queued DB messages, then pushes in real-time
- send_message: pushes to WS if connected, falls back to DB queue
- Auto-cleanup on disconnect
- WS accepts both binary and JSON text frames for sending
Web client:
- Replaces 2-second HTTP polling with persistent WebSocket
- Auto-reconnects on disconnect (3-second backoff)
- Sends via WS when connected, HTTP fallback
- Messages arrive instantly (no polling delay)
- "Real-time connection established" shown on connect
HTTP polling still works:
- CLI recv command uses HTTP (unchanged)
- Web falls back to HTTP if WS fails
- Mules/scripts can still use HTTP API
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Each peer gets a stable color from a 12-color palette based on
their fingerprint/alias hash. Self messages stay green.
No more same-color for different users.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: web client's bundle included OTPKs, so X3DH initiate()
did 4 DH ops (DH4 with OTPK). But decrypt_wire_message() called
respond() with None for OTPK, doing only 3 DH ops.
Different DH concat → different shared secret → decrypt fails.
Fix: web client bundles have one_time_pre_key: None.
initiate() skips DH4 when no OTPK present.
respond() also skips DH4 with None.
Both sides now do exactly 3 DH ops → shared secrets match.
OTPKs are an anti-replay optimization, not required for E2E.
Will add OTPK support to web client in Phase 2 with proper
server-side OTPK storage and consumption tracking.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Version shown on chat load (v0.0.2)
- Self-test now does step-by-step: X3DH shared secret comparison,
then manual ratchet init + decrypt (not via decrypt_wire_message)
- Shows: rng output, shared_match, alice/bob shared secrets, decrypt result
- This isolates whether X3DH or ratchet or AEAD fails
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: WASM was regenerating random pre-keys on every call to
decrypt_wire_message, instead of using the SPK that was registered
with the server. CLI sender encrypts to the registered SPK, but
WASM was trying to decrypt with a different random key.
Fix:
- WasmIdentity now stores spk_secret_bytes internally
- SPK secret persisted to localStorage as 'wz-spk'
- On load: restored from localStorage, not regenerated
- bundle_bytes() uses stored SPK secret (cached, deterministic)
- decrypt_wire_message() takes spk_secret_hex parameter
- Web UI passes stored SPK to all decrypt calls
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
warzone-wasm crate:
- Compiles warzone-protocol to WebAssembly via wasm-pack
- Exposes WasmIdentity, WasmSession, decrypt_wire_message to JS
- Same X25519 + ChaCha20-Poly1305 + X3DH + Double Ratchet as CLI
- 344KB WASM binary (optimized with wasm-opt)
WireMessage moved to warzone-protocol:
- Shared type used by CLI client, WASM bridge, and TUI
- Guarantees identical bincode serialization across all clients
Web client rewritten:
- Loads WASM module on startup (/wasm/warzone_wasm.js)
- Identity: WasmIdentity generates same key types as CLI
- Registration: sends bincode PreKeyBundle (same format as CLI)
- Encrypt: WasmSession.encrypt/encrypt_key_exchange
- Decrypt: decrypt_wire_message (handles KeyExchange + Message)
- Sessions persisted in localStorage (base64 ratchet state)
- Groups: per-member WASM encryption (interop with CLI members)
Server routes:
- GET /wasm/warzone_wasm.js — serves WASM JS glue
- GET /wasm/warzone_wasm_bg.wasm — serves WASM binary
- Both embedded at compile time via include_str!/include_bytes!
Web ↔ CLI interop now works:
- Same key exchange (X3DH with X25519)
- Same ratchet (Double Ratchet with ChaCha20-Poly1305)
- Same wire format (bincode WireMessage)
- Web user can message CLI user and vice versa
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Aliases now have a lifecycle:
- 365-day TTL from last activity (send/receive/renew)
- 30-day grace period after expiry (only recovery key can reclaim)
- After grace: anyone can register the alias
- Recovery key generated on first registration, rotated on recovery
- Auto-renew on activity via POST /v1/alias/renew
New endpoints:
- POST /v1/alias/recover {alias, recovery_key, new_fingerprint}
Reclaim alias with recovery key, even if expired. Works across
identity changes (new seed → new fingerprint, same alias).
Recovery key is rotated on each recovery.
- POST /v1/alias/renew {fingerprint}
Heartbeat — resets TTL. Returns days until expiry.
Resolve now returns expiry info:
- GET /v1/alias/resolve/:name → includes expires_in_days, expired flag
- GET /v1/alias/list → includes expiry status per alias
Phase 2: DNS automation — separate DNS authority manages parent zone,
servers update delegated records via API. Recovery key maps to DNS
record ownership for out-of-band reclamation.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
