The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/clip_video.py executes the ffmpeg binary via subprocess.run to process video files.
The command is constructed using a list of arguments, which prevents shell injection vulnerabilities by avoiding the use of a shell for execution.
This behavior is the primary intended functionality of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided subtitle files which may contain malicious instructions designed to influence the AI's analysis.
Ingestion points: Subtitle content is read from external .vtt and .srt files in scripts/clip_video.py within the load_subtitles function.
Boundary markers: The instructions lack explicit delimiters or warnings to the AI to ignore instructions embedded within the subtitle data.
Capability inventory: The skill possesses the capability to write to the local filesystem and execute the ffmpeg utility via the Python script.
Sanitization: Subtitle text is passed to the AI for semantic analysis without prior sanitization or filtering to remove potential prompt injection attempts.

semantic-video-clipper