The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses instructional language to guide the agent in generating new content. No bypass markers, override commands, or attempts to extract system prompts were found.
[DATA_EXFILTRATION]: No network operations or access to sensitive file paths (such as SSH keys or credentials) are present.
[REMOTE_CODE_EXECUTION]: The skill defines a folder structure that includes a scripts directory for helper scripts, but it does not provide any logic to download or execute external code from remote sources.
[COMMAND_EXECUTION]: While the skill instructions mention using specific bash commands for low-freedom tasks in generated skills, this skill itself does not execute any commands or sub-processes.

creating-skills