The Agent Skills Directory

[SAFE]: The skill generates interactive bash scripts designed to facilitate manual setup tasks. The template.sh file contains utility functions for user prompts, URL opening, and writing configuration files (.env).
[COMMAND_EXECUTION]: The script uses standard system commands (tput, gh, xdg-open, open) to manage the CLI interface and interact with local tools. These are legitimate uses within the scope of a setup wizard.
[DATA_EXPOSURE_AND_EXFILTRATION]: While the script handles sensitive data like API keys, it does so by prompting the user (ask_secret) and saving the values to local files (.env) or the user's own GitHub repository via the gh CLI. It does not send data to any external or third-party servers.
[CREDENTIALS_UNSAFE]: The template includes placeholders and examples (e.g., STRIPE_SECRET_KEY), but does not contain any hardcoded credentials. It correctly implements ask_secret to prevent secrets from being echoed to the terminal and sets file permissions of .env to 0600 to restrict access.
[REMOTE_CODE_EXECUTION]: The skill instructions advise the user on how to verify and run the generated script locally (bash -n, chmod +x). It does not download or execute remote scripts.

wizard