This skill provides a rigorous framework for creating 'run' skills for any software project. Instead of just providing a README, it forces the agent to build a programmatic 'driver' (e.g., via chromium-cli, tmux, or curl) that allows future agents to actually interact with the running application, take screenshots, and verify functionality.
Use this skill when you need to set up a project's runtime environment, create instructions for launching the app, or develop an interaction harness to verify build and run steps in a headless container.
Primarily designed for Claude Code and other agents capable of executing shell commands and managing tmux/xvfb sessions.
This skill has not been reviewed by our automated audit pipeline yet.