The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a Python script (scripts/video_frames.py) to execute ffmpeg and ffprobe for media processing. These executions are performed using subprocess.run with list-based arguments and no shell invocation (shell=False), which is a secure implementation for the skill's core functionality. Input data derived from video metadata is properly sanitized through type conversion.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing untrusted external recordings and codebases. While this presents an inherent surface for indirect prompt injection, the skill incorporates explicit instructions for evidence classification (Observed, Corroborated, and Inferred) to help the agent maintain objective analysis and avoid being misled by untrusted content.
[DATA_EXFILTRATION]: The instructions include a safety directive to keep recordings and derived evidence local and uncommitted by default, protecting user data from unintended exposure during the analysis process.

analyze-ui-video