The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/init_skill.py utility creates local directories and template files using the pathlib library. It also applies executable permissions to generated scripts using chmod.
[COMMAND_EXECUTION]: The scripts/package_skill.py script reads the contents of the skill directory and archives them into a ZIP file using the zipfile module.
[PROMPT_INJECTION]: As a tool for generating agent instructions, the skill includes scripts/quick_validate.py which implements basic sanitization by checking for angle brackets in descriptions and enforcing strict kebab-case naming conventions for skill identifiers.
[SAFE]: The Python implementation uses yaml.safe_load() to prevent unsafe deserialization during metadata validation and does not perform any network operations or external downloads.

skill-creator