The Agent Skills Directory

COMMAND_EXECUTION (SAFE): The skill includes Python scripts (scripts/init_skill.py and scripts/validate_skill.py) that perform local directory creation and file structure validation. These scripts utilize the standard Python pathlib and argparse libraries and do not execute external or shell-level commands that could be exploited for arbitrary execution.
PROMPT_INJECTION (SAFE): The core instructions and the provided skill templates follow standard formatting and do not contain patterns intended to bypass agent safety filters, extract system prompts, or override core agent behavior.
DATA_EXFILTRATION (SAFE): No network requests, hardcoded credentials, or access to sensitive local configuration files (such as SSH keys or environment variables) were found. All data processing is confined to the skill's local development lifecycle.

skill-creator