The Agent Skills Directory

[SAFE]: The skill implements a privacy-preserving telemetry initializer in SKILL.md that explicitly excludes sensitive data like credentials, business data, or URLs.
[COMMAND_EXECUTION]: The skill uses subprocess.run in tests/test_scripts.py to execute its own internal validation scripts (audit_edit_boundaries.py and validate_delivery.py). This is a legitimate use for testing its own functionality and does not involve external or untrusted inputs.
[DATA_EXFILTRATION]: No patterns of unauthorized data exfiltration were found. The telemetry_init.py mentioned is described as privacy-preserving and opt-in, specifically avoiding logging of user content or credentials.
[PROMPT_INJECTION]: The instructions in SKILL.md emphasize grounding analysis in authorized sources and explicitly require user approval and readback for any action involving publishing, scheduling, or quota spending, which acts as a safeguard against autonomous agent abuse.

video-content-engine