cli-agent-runner

What problem does it solve?

This Skill simplifies invoking long-running Claude Code tasks by exposing a simple, named session runner. It hides the complexity of session IDs, supports background execution with polling, and standardizes result extraction so you can focus on the task rather than orchestration.

Core Features & Use Cases

• Named sessions: Create, resume, and monitor Claude Code work using memorable session names instead of IDs.
• Background execution & polling: Run tasks in the background and check for completion without blocking your workflow.
• Automatic result extraction: Cleanly capture and present the task result on stdout while sending errors to stderr.
• Agent support: Optional agent definitions to tailor behavior for specialized tasks.

Quick Start

Start a new session named architect with the prompt "Design a high-level architecture document for a user authentication system".

1. Determine the path to the script: Identify the root folder of the plugin skill and append the script name: <path-to-skill-root-folder>/cli-agent-runner.sh
2. Start a new session: <absolute-path-to-cli-agent-runner.sh> new architect -p "Design a high-level architecture document for a user authentication system"
3. Later, resume or check status: <absolute-path-to-cli-agent-runner.sh> resume architect -p "Add API endpoint specs"