The Agent Skills Directory

[SAFE]: The skill facilitates the creation of interactive bash scripts. While these scripts handle sensitive information such as API keys and secrets, they use best-practice patterns:
Secrets are collected via hidden input (ask_secret).
Values are persisted locally to standard .env files using idempotent logic.
GitHub secrets are updated using the official gh CLI tool, ensuring secure transmission to the user's own repository.
The network operations (opening URLs) are performed via standard cross-platform system openers (xdg-open, open, explorer.exe) directed at official service dashboards (e.g., Stripe) determined by the user or developer.
The skill uses a disable-model-invocation: true flag, which limits the agent's autonomy and ensures the logic remains focused on the provided template.

wizard