The Agent Skills Directory

[SAFE]: The skill acts as a scaffolding tool for creating other agent skills. It utilizes local Python scripts for directory initialization and validation without making external network requests.
[COMMAND_EXECUTION]: The skill invokes its own internal Python scripts to perform structured tasks like validation and packaging. These executions are confined to the skill's own directory and use environment-variable-derived paths for location-agnostic execution.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it generates instructions based on user input. This is a characteristic of the skill's intended purpose as a generator and is mitigated by the inclusion of validation scripts and structural templates.
[SAFE]: The skill documents the use of dynamic context injection for developer workflows, providing instructions on how to legitimately integrate shell command output (like git status) into the agent context.

create-skill