The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts (scripts/init_skill.py, scripts/package_skill.py) that perform local file system operations, such as creating directories, writing template files, and zipping skill folders. These operations are standard for a project scaffolding and packaging tool.
[PRIVILEGE_ESCALATION]: The scripts/init_skill.py script programmatically modifies file permissions using chmod(0o755) on a dynamically created example script. This behavior is intended to make the generated template immediately functional for the developer but involves changing execution privileges.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-defined skill names and file paths, presenting a potential surface for indirect injection if inputs are not sanitized.
Ingestion points: Command-line arguments for skill name and destination path in scripts/init_skill.py.
Boundary markers: Absent; input strings are directly interpolated into file system paths and SKILL.md templates.
Capability inventory: Local file system write and directory creation in scripts/init_skill.py; ZIP archival in scripts/package_skill.py.
Sanitization: Kebab-case regex validation (^[a-z0-9-]+$) is implemented in scripts/quick_validate.py to enforce safe naming conventions for the skill directory.

skill-creator