LaunchKit · 2026
Back to Skills

claude-code-pro

Token-efficient Claude Code workflow.

0
462 downloads
by @swaylq

Setup & Installation

openclaw skills install @swaylq/claude-code-pro

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install claude-code-pro

Version History

v1.1.0Feb 24, 2026 - Version note by skill developer:

Fix bracketed paste bug, remove jq dependency, add task guard, add README, add safety warnings

What This Skill Does

A tmux-based workflow for running Claude Code tasks in background sessions using completion callbacks instead of polling. Tasks signal their own completion via an event command, eliminating repeated output reads. Suited for multi-file coding work that needs to run unattended.

Uses completion callbacks instead of polling, cutting supervision token usage by 80-97% compared to skills that read terminal output every 30-60 seconds.

When to use it

  • Refactoring an auth module to JWT while continuing other work
  • Implementing a new feature across 5+ files in the background
  • Running a multi-file code migration without active supervision
  • Confirming a long-running background task completed with a single check
  • Deciding whether to spawn a session or just edit files directly

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: Refactor the auth module to use JWT across multiple files

AGENT
  1. 1Check that the task involves 3+ files, confirming Claude Code is appropriate
  2. 2Run start.sh with the task description and a completion callback command appended
  3. 3Wait for the openclaw system event notification signaling task completion
  4. 4Run monitor.sh --lines 50 to confirm the result
  5. 5Run stop.sh to clean up the tmux session
OUTPUT

Auth module refactored to JWT; session closed after one monitoring read

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

Claude Code CLI (claude)