chore: align setup with EJClaw dual-service architecture

.env.example

@@ -1,8 +1,44 @@
+# ===== EJClaw Environment Configuration =====
+# Copy this file to .env and fill in the values.
+# For dual-service (Codex), also create .env.codex with DISCORD_BOT_TOKEN for the Codex bot.
 
+# --- Required ---
+DISCORD_BOT_TOKEN=                    # Discord bot token (Claude bot)
+CLAUDE_CODE_OAUTH_TOKEN=              # Claude Code OAuth token (primary account)
 
-# Added by skill
-DISCORD_BOT_TOKEN=
+# --- Multi-account rotation (recommended) ---
+CLAUDE_CODE_OAUTH_TOKENS=             # Comma-separated tokens for auto-rotation on 429
+                                      # e.g. token1,token2
 
-# Codex support (optional — only needed for Codex agent type)
-OPENAI_API_KEY=
-# CODEX_OPENAI_API_KEY=  # Use this to override OPENAI_API_KEY for Codex only
+# --- Codex support (optional — needed for ejclaw-codex service) ---
+OPENAI_API_KEY=                       # OpenAI API key for Codex
+# CODEX_OPENAI_API_KEY=              # Override OPENAI_API_KEY for Codex only
+# CODEX_MODEL=gpt-5.4                # Codex model
+# CODEX_EFFORT=xhigh                 # Codex reasoning effort
+
+# --- Bot identity ---
+ASSISTANT_NAME=claude                 # Trigger name (@claude)
+# SERVICE_ID=claude                   # Service identifier (defaults to ASSISTANT_NAME slug)
+# SERVICE_AGENT_TYPE=claude-code      # Agent type: claude-code or codex
+
+# --- Voice transcription ---
+GROQ_API_KEY=                         # Groq Whisper (primary, free: console.groq.com)
+# OPENAI_API_KEY also used as Whisper fallback
+
+# --- Provider fallback (optional — Claude 429 → alternative) ---
+# FALLBACK_PROVIDER_NAME=kimi
+# FALLBACK_BASE_URL=https://api.kimi.com/coding
+# FALLBACK_AUTH_TOKEN=
+# FALLBACK_MODEL=kimi-k2.5
+# FALLBACK_COOLDOWN_MS=600000
+
+# --- Usage dashboard (optional) ---
+# STATUS_CHANNEL_ID=                  # Discord channel for live status updates
+# USAGE_DASHBOARD=true                # Enable usage dashboard
+# STATUS_SHOW_ROOMS=true              # Show room status in dashboard
+# STATUS_SHOW_ROOM_DETAILS=false      # Show detailed room info
+
+# --- Advanced ---
+# MAX_CONCURRENT_AGENTS=5             # Max parallel agent processes
+# AGENT_TIMEOUT=1800000               # Agent timeout in ms (default: 30min)
+# SESSION_COMMAND_ALLOWED_SENDERS=    # Comma-separated Discord user IDs for session commands