@thesammykins/tether 1.0.0 → 1.2.0

package/.env.example

@@ -12,6 +12,20 @@ AGENT_TYPE=claude
 # Optional - Working directory for agent sessions
 CLAUDE_WORKING_DIR=/path/to/your/project
 
+# Optional - HTTP API server configuration
+TETHER_API_HOST=127.0.0.1
+# Bind address for the HTTP API server (default: 127.0.0.1)
+# Use 0.0.0.0 to allow external connections
+
+TETHER_API_PORT=2643
+# Port for the HTTP API server (default: 2643)
+
+API_TOKEN=
+# Optional - API authentication token
+# If set, all API requests (except /health) must include:
+# Authorization: Bearer <token>
+# If not set, no authentication is required (backward compatible)
+
 # Optional - Direct Messages (disabled by default)
 ENABLE_DMS=false
 # Set to "true" to allow the bot to respond to DMs.

package/README.md

@@ -1,265 +1,53 @@
 # @thesammykins/tether
 
+[![npm version](https://img.shields.io/npm/v/@thesammykins/tether)](https://www.npmjs.com/package/@thesammykins/tether)
+[![license](https://img.shields.io/npm/l/@thesammykins/tether)](./LICENSE)
+
 Discord bot that bridges messages to AI coding agents. Supports Claude Code, OpenCode, and Codex CLI.
 
 Fork of [cord](https://github.com/alexknowshtml/cord) with multi-agent support and enhanced features.
 
 ## Features
 
-- **Multi-agent support** — Claude Code, OpenCode, Codex (switch via `AGENT_TYPE` env var)
-- **Direct messages** — Chat with the bot privately via DMs (opt-in)
+- **Multi-agent support** — Claude Code, OpenCode, Codex (switch via config)
+- **BRB mode** — Agent asks questions via Discord buttons when you're away
+- **Direct messages** — Chat with the bot privately via DMs
+- **Encrypted config** — `tether config` stores tokens with AES-256-GCM encryption
 - **Access control** — User, role, and channel allowlists
 - **Rate limiting** — Per-user sliding window
 - **Session management** — Turn limits, duration limits, pause/resume
 - **Smart threads** — Auto-naming, channel context for new conversations
-- **Acknowledgment** — 👀 reaction on incoming messages
 - **Resilient** — Exponential backoff on connection failures
 
 ## Quick Start
 
-### Prerequisites
-
-- [Bun](https://bun.sh) runtime
-- [Redis](https://redis.io) (for BullMQ job queue)
-- A Discord bot token ([setup guide below](#discord-bot-setup))
-- An AI agent CLI installed on PATH (`claude`, `opencode`, or `codex`)
-
-### Install
-
-```bash
-git clone https://github.com/thesammykins/tether.git
-cd tether
-bun install
-```
-
-### Configure
-
-```bash
-cp .env.example .env
-# Edit .env with your DISCORD_BOT_TOKEN and settings
-```
-
-### Run
-
 ```bash
-bun run start        # Start bot + worker via CLI
-# or run separately:
-bun run bot          # Discord gateway
-bun run worker       # Job processor
-```
-
----
-
-## Discord Bot Setup
-
-Step-by-step guide to creating and configuring your Discord bot.
-
-### 1. Create a Discord Application
-
-1. Go to the [Discord Developer Portal](https://discord.com/developers/applications)
-2. Click **New Application** and give it a name (e.g. "Tether")
-3. Note the **Application ID** — you'll need it for the invite link
-
-### 2. Create the Bot User
-
-1. In your application, go to the **Bot** tab in the left sidebar
-2. Click **Add Bot** (if not already created)
-3. Under **Token**, click **Reset Token** and copy it — this is your `DISCORD_BOT_TOKEN`
-4. **Store this token securely** — you won't be able to see it again
-
-### 3. Configure Bot Permissions
-
-Under the **Bot** tab, configure these settings:
-
-**Privileged Gateway Intents** (toggle ON):
-
-| Intent | Why |
-|--------|-----|
-| **Message Content Intent** | Required to read message text (not just metadata) |
-
-> Without Message Content Intent enabled, the bot will connect but never see message contents.
-
-**Bot Permissions** — the bot needs these permissions in your server:
-
-| Permission | Why |
-|------------|-----|
-| Send Messages | Reply to users |
-| Create Public Threads | Create conversation threads |
-| Send Messages in Threads | Respond in threads |
-| Manage Threads | Rename threads |
-| Read Message History | Fetch channel context for new conversations |
-| Add Reactions | 👀 acknowledgment emoji |
-| Use Slash Commands | `/cord config` command |
-
-### 4. Generate the Invite Link
-
-1. Go to the **OAuth2** tab → **URL Generator**
-2. Under **Scopes**, select:
-   - `bot`
-   - `applications.commands`
-3. Under **Bot Permissions**, select the permissions listed above, or use the permission integer `326417588288` which includes all required permissions
-4. Copy the generated URL and open it in your browser
-5. Select the server you want to add the bot to and click **Authorize**
-
-### 5. Enable DMs (Optional)
-
-If you want users to be able to DM the bot directly:
-
-1. In `.env`, set `ENABLE_DMS=true`
-2. In the Discord Developer Portal → **Bot** tab, make sure **Allow DMs** is not disabled
-
-DM behavior:
-- Any message sent to the bot in DMs starts or continues a conversation
-- No `@mention` needed — every DM is treated as a prompt
-- Sessions persist per-user until manually reset
-- Send `!reset` in DMs to start a fresh session
-- Only `ALLOWED_USERS` is checked for DMs (roles and channels don't apply)
-
-### 6. Start the Bot
-
-```bash
-# Make sure Redis is running
-redis-server
-
-# Start tether
-bun run start
-```
-
-You should see:
-```
-[bot] Connecting to Discord gateway (attempt 1)...
-[bot] Logged in as YourBot#1234
-[worker] Worker started, waiting for jobs...
-```
-
-### 7. Test It
-
-In your Discord server, `@mention` the bot in any channel:
+# Install
+bun add -g @thesammykins/tether
 
-```
-@Tether what time is it?
-```
-
-The bot will:
-1. React with 👀
-2. Create a thread with an auto-generated name
-3. Post a "Processing..." status message
-4. Forward the prompt to your AI agent
-5. Post the response in the thread
-
----
-
-## Configuration
-
-| Variable | Default | Description |
-|----------|---------|-------------|
-| `DISCORD_BOT_TOKEN` | required | Discord bot token |
-| `AGENT_TYPE` | `claude` | Agent backend: `claude`, `opencode`, `codex` |
-| `ENABLE_DMS` | `false` | Allow bot to respond to direct messages |
-| `REDIS_HOST` | `localhost` | Redis host |
-| `REDIS_PORT` | `6379` | Redis port |
-| `ALLOWED_USERS` | (empty=all) | Comma-separated Discord user IDs |
-| `ALLOWED_ROLES` | (empty=all) | Comma-separated role IDs (guild only) |
-| `ALLOWED_CHANNELS` | (empty=all) | Comma-separated channel IDs (guild only) |
-| `RATE_LIMIT_REQUESTS` | `5` | Max requests per window |
-| `RATE_LIMIT_WINDOW_MS` | `60000` | Rate limit window (ms) |
-| `MAX_TURNS_PER_SESSION` | `50` | Max turns per thread/DM session |
-| `MAX_SESSION_DURATION_MS` | `3600000` | Max session duration (ms) |
-| `CLAUDE_WORKING_DIR` | cwd | Default working directory for agents |
-| `DB_PATH` | `./data/threads.db` | SQLite database path |
-| `CORD_ALLOWED_DIRS` | (empty=any) | Comma-separated allowed working directories |
-
-### Finding Discord IDs
-
-To get user, role, or channel IDs:
-1. Enable **Developer Mode** in Discord: Settings → App Settings → Advanced → Developer Mode
-2. Right-click a user, role, or channel and select **Copy ID**
-
-## Agent CLI Requirements
-
-Whichever agent you choose must be installed and available on PATH:
-- **Claude Code**: `claude` CLI — [Install guide](https://docs.anthropic.com/en/docs/claude-code)
-- **OpenCode**: `opencode` CLI
-- **Codex**: `codex` CLI
-
-## Architecture
+# Configure
+tether config set DISCORD_BOT_TOKEN     # paste token (hidden input)
+tether config set AGENT_TYPE claude      # or: opencode, codex
 
-Tether uses a queue-based architecture for reliable message processing:
-
-```
-Discord Bot  →  BullMQ Queue  →  Agent Worker
- (gateway)       (Redis)          (spawns CLI)
+# Start Redis + Tether
+redis-server &
+tether start
 ```
 
-### Key Components
-
-- **Bot** (`src/bot.ts`) — Discord gateway with middleware pipeline
-- **Worker** (`src/worker.ts`) — Job processor using adapter registry
-- **Adapters** (`src/adapters/`) — AgentAdapter interface for Claude/OpenCode/Codex
-- **Middleware** (`src/middleware/`) — Allowlist, rate-limiter
-- **Features** (`src/features/`) — Ack, channel-context, thread-naming, session-limits, pause-resume
-- **Queue** (`src/queue.ts`) — BullMQ job queue
-- **DB** (`src/db.ts`) — SQLite for thread→session mapping
-
-### Message Flow
-
-**Guild channels:**
-1. User `@mentions` bot in Discord channel
-2. Bot checks allowlist → rate limiter → pause state
-3. Bot creates thread with auto-generated name
-4. Bot adds job to BullMQ queue
-5. Worker pulls job and spawns agent adapter
-6. Adapter runs CLI (`claude`/`opencode`/`codex`) with prompt
-7. Worker posts response back to Discord thread
-
-**Direct messages:**
-1. User sends message to bot in DMs (no `@mention` needed)
-2. Bot checks user allowlist → rate limiter
-3. Bot creates or resumes session for this user
-4. Bot adds job to BullMQ queue
-5. Worker posts response back to the DM channel
-
-### Middleware Pipeline
-
-Messages pass through middleware in this order:
-
-1. **Allowlist** — Block unauthorized users/channels/roles (DMs: user-only check)
-2. **Rate Limiter** — Sliding window per-user rate limit
-3. **Pause/Resume** — Hold messages if thread is paused (guild threads only)
-
-### Session Management
-
-- **Turn limits** — Max messages per thread/DM session (prevents infinite loops)
-- **Duration limits** — Max session lifetime
-- **Pause/Resume** — Type `pause` to hold messages, `resume` to continue (threads only)
-- **DM Reset** — Type `!reset` in DMs to start a new session
-- **Auto-completion** — React with ✅ on last message to mark "Done" (threads only)
+Then `@mention` the bot in your Discord server.
 
-## Working Directory Configuration
+## Documentation
 
-Tether supports per-channel working directories so agents operate in the correct project context.
-
-**Channel-level configuration** (persists for all conversations in that channel):
-```
-/cord config dir ~/Code/myproject
-```
-
-**Per-message override** (one-time, just for this conversation):
-```
-@bot [/other/project] what files are here?
-```
-
-**Fallback chain**: Message override → Channel config → `CLAUDE_WORKING_DIR` env → `process.cwd()`
-
-### Security: Directory Allowlist
-
-For multi-user deployments, restrict which directories users can access:
-
-```bash
-CORD_ALLOWED_DIRS=/home/projects,/var/code
-```
-
-If not set, any existing directory is allowed (backward compatible). When set, paths outside the allowlist are rejected.
+| Guide | Description |
+|-------|-------------|
+| **[Installation](docs/installation.md)** | Prerequisites, install methods, quick setup |
+| **[Discord Setup](docs/discord-setup.md)** | Bot creation, permissions, intents, invite link |
+| **[Agent Setup](docs/agents.md)** | Claude Code / OpenCode / Codex install and auth |
+| **[Configuration](docs/configuration.md)** | All config keys, encrypted secrets, resolution chain |
+| **[CLI Reference](docs/cli.md)** | Every `tether` command with examples |
+| **[Architecture](docs/architecture.md)** | System design, message flow, middleware pipeline |
+| **[Troubleshooting](docs/troubleshooting.md)** | Common problems and solutions |
+| **[HTTP API](skills/tether/HTTP-API.md)** | REST API for external scripts and webhooks |
 
 ## Testing
 
@@ -268,45 +56,11 @@ bun test                     # Run all tests
 bun test tests/adapters/     # Adapter tests only
 bun test tests/middleware/   # Middleware tests only
 bun test tests/features/     # Feature tests only
-bun test tests/integration/  # Integration tests only
 ```
 
-## CLI Commands
-
-See [skills/tether/SKILL.md](./skills/tether/SKILL.md) for full CLI documentation.
-
-Quick reference:
-- `tether send <channel> "message"` — Send text message
-- `tether embed <channel> "text" --title "T"` — Send formatted embed
-- `tether file <channel> ./file.txt` — Send file attachment
-- `tether buttons <channel> "prompt" --button label="Yes" id="yes"` — Send interactive buttons
-- `tether state <channel> <msgId> done` — Update status message
-- `tether dm <user-id> "message"` — Send a DM to a user (proactive outreach)
-- `tether dm <user-id> --embed "text" --title "T"` — Send an embed DM
-- `tether dm <user-id> --file ./report.md` — Send a file via DM
-
-## HTTP API
-
-Tether exposes an HTTP API on port 2643 for external scripts and webhooks.
-
-See [skills/tether/HTTP-API.md](./skills/tether/HTTP-API.md) for API documentation.
-
-## Troubleshooting
-
-| Problem | Solution |
-|---------|----------|
-| Bot connects but doesn't respond | Enable **Message Content Intent** in Developer Portal → Bot tab |
-| "TokenInvalid" error | Regenerate your bot token in the Developer Portal |
-| "DisallowedIntents" error | Enable the required intents in Developer Portal → Bot tab |
-| Bot doesn't receive DMs | Set `ENABLE_DMS=true` in `.env` |
-| "Rate limit exceeded" | Adjust `RATE_LIMIT_REQUESTS` / `RATE_LIMIT_WINDOW_MS` |
-| Agent command not found | Ensure `claude`/`opencode`/`codex` is installed and on PATH |
-| Redis connection refused | Start Redis: `redis-server` |
-| Bot can't create threads | Check bot has **Create Public Threads** permission in your server |
-
 ## Privacy
 
-Tether stores only thread-to-session mappings and channel configuration (working directories). **No message content, user data, or conversation history is persisted.** Messages pass through the Redis queue transiently and are discarded after processing.
+No message content or user data is stored. Only thread-to-session mappings and channel config persist. Messages pass through Redis transiently and are discarded after processing.
 
 ## License