Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Runs long-running shell tasks without hitting HTTP timeout limits. Agents call start, execute commands, then push the result back to the active session when done. Auto-detects the session via OpenClaw or Clawdbot CLI.
Eliminates empty response errors from HTTP timeouts by decoupling task execution from the request/response cycle.
When to use it
- Counting or scanning files across large codebases
- Running multi-step build or test pipelines
- Fetching data from slow or unpredictable external APIs
- Analyzing large log files without dropping the connection
- Executing background jobs that exceed 5-second response limits
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Count all TypeScript files in this project
- 1Call async-task start "Counting TypeScript files..." to acknowledge immediately
- 2Run find . -name "*.ts" | wc -l to count the files
- 3Call async-task done "Found $count TypeScript files" to push result to the active session
User receives the file count pushed directly to their session