The Agent Skills Directory

[SAFE]: The skill is a Markdown-based reference document and does not contain any executable scripts or hidden logic.\n- [COMMAND_EXECUTION]: Provides documented shell command examples for interacting with the Codex CLI, including standard data processing with jq and grep.\n- [SAFE]: Uses clear placeholders for API key instructions (e.g., <api-key>), following best practices to avoid hardcoding actual credentials.\n- [SAFE]: Permission models and high-privilege flags such as --sandbox danger-full-access are explicitly documented with appropriate warnings regarding their usage in isolated environments.

codex-reference