The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/render.py executes system commands using the list-based subprocess.run interface. This pattern is secure against shell injection because arguments are not interpreted by a shell. Numerical inputs like dimensions and duration are strictly validated via argparse types.\n- [DATA_EXFILTRATION]: No network exfiltration or sensitive file access patterns were detected. The skill only interacts with local files for image input and video output.\n- [PROMPT_INJECTION]: The SKILL.md instructions are purely technical and provide no evidence of attempts to bypass AI safety guards or override instructions.\n- [REMOTE_CODE_EXECUTION]: The skill does not perform remote downloads or dynamic execution of untrusted code. It relies on standard libraries and pre-installed system utilities.

ken-burns-clip