name: ClawConfig
description: Moltbot/Clawdbot Telegram bot configuration and maintenance. USE WHEN moltbot, clawdbot, telegram bot, bot config, bot debug, bot deploy, gateway status, bot not responding, bot maintenance, cloudflare sandbox.

ClawConfig

Configuration, deployment, debugging, and maintenance for the Moltbot/Clawdbot Telegram bot running on Cloudflare Sandbox.

Customization

Before executing, check for user customizations at:
~/.claude/skills/PAI/USER/SKILLCUSTOMIZATIONS/ClawConfig/

If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.

🚨 MANDATORY: Voice Notification (REQUIRED BEFORE ANY ACTION)

You MUST send this notification BEFORE doing anything else when this skill is invoked.

1. Send voice notification:
   bash curl -s -X POST http://localhost:8888/notify \ -H "Content-Type: application/json" \ -d '{"message": "Running the WORKFLOWNAME workflow in the ClawConfig skill to ACTION"}' \ > /dev/null 2>&1 &

2. Output text notification:
   Running the **WorkflowName** workflow in the **ClawConfig** skill to ACTION...

This is not optional. Execute this curl command immediately upon skill invocation.

Workflow Routing

Critical Paths

Quick Reference

Architecture:
- Cloudflare Worker → Cloudflare Sandbox → Moltbot Container → Telegram API
- Config hash in /tmp/gateway-config-hash tracks version
- Gateway auto-restarts when config hash changes

Key Commands:

# Build
cd /root/moltworker && npm run build

# Deploy (requires CLOUDFLARE_API_TOKEN)
CLOUDFLARE_API_TOKEN=xxx npx wrangler deploy

# Tail logs
CLOUDFLARE_API_TOKEN=xxx npx wrangler tail --format=pretty

Debug Endpoints (require CF Access or DEBUG_ROUTES=true):
- /debug/version-check - Config hash comparison
- /debug/processes - Container process list
- /debug/logs - Gateway stdout/stderr
- /debug/container-config - Moltbot JSON config
- /debug-public - Public debug info (no auth)

Examples

Example 1: Deploy after code changes

User: "deploy the bot changes"
→ Invokes Deploy workflow
→ Runs npm build
→ Executes wrangler deploy with API token
→ Reports version ID and confirms deployment

Example 2: Bot not responding to messages

User: "telegram bot isn't responding"
→ Invokes Debug workflow
→ Checks wrangler tail for errors
→ Verifies config hash status
→ Checks for CF Access redirect issues
→ Identifies root cause and fixes

Example 3: Check current gateway status

User: "is the gateway running?"
→ Invokes Status workflow
→ Hits /debug/version-check endpoint
→ Reports gateway status, config hash match, process ID

Example 4: Change the default model

User: "switch to Sonnet instead of Opus"
→ Invokes Config workflow
→ Edits start-moltbot.sh default model
→ Deploys changes
→ Gateway auto-restarts with new model

Common Issues