Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Sync latest shared skills update (9ed51f1): async launch semantics, testing protocol docs, and wake/routing hardening refinements. WhatsApp E2E smoke test passed.
What This Skill Does
Runs Claude Code in a detached background process and delivers results to Telegram or WhatsApp when complete. Supports coding, codebase research, file analysis, content generation, and multi-step automations. Includes strict Telegram thread-safe routing, silent heartbeat pings during execution, and iterative continuation mode.
Runs Claude Code on a Max subscription with zero API token cost while the main agent session stays free for other work.
When to use it
- Refactoring a codebase overnight and receiving the result on mobile
- Running competitive research while away from the terminal
- Analyzing production logs and getting a structured summary in a Telegram thread
- Generating project documentation with WhatsApp delivery
- Running multi-step automation workflows without monitoring the process
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Refactor the auth module to use JWT and notify me in my Telegram thread
- 1Saves task prompt to /tmp/cc-prompt.txt to avoid shell quoting issues
- 2Validates Telegram thread routing with --validate-only flag before launch
- 3Launches run-task.py via nohup with the thread session key and a 900s timeout
- 4Background process runs Claude Code and sends silent heartbeat pings to the thread every 60 seconds
- 5On completion, delivers the result to the Telegram thread and wakes the agent for a visible continuation summary turn
Refactored auth module with JWT support delivered as a Telegram thread message, followed by an agent continuation turn summarizing changes and next steps
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.