The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the use of an external command-line interface (codex) to perform autonomous operations. It correctly advises on the use of sandboxing flags like --sandbox workspace-write (to allow file modifications) and read-only (for research tasks) to enforce the principle of least privilege.\n- [SAFE]: No malicious patterns such as obfuscation, credential harvesting, or unauthorized remote code execution were found. The skill relies on a pre-installed local utility and instructs the agent to verify the tool's existence before attempting to use it.

use-codex