The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements safety guidelines in SKILL.md to refuse harmful or violent content. The implementation in src/index.js utilizes a guardRequest function to sanitize and validate user input before processing.
[COMMAND_EXECUTION]: The script scripts/validate_creative_artifacts.py uses subprocess.run to call a local validation utility. This execution is limited to internal vendor scripts located in the relative shared directory and does not expose the system to arbitrary command execution from untrusted sources.
[DATA_EXFILTRATION]: Input data is processed locally to generate creative text artifacts. There are no network operations, hardcoded credentials, or attempts to access sensitive system files (e.g., SSH keys or environment variables).

action-master-kinetic-director