The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The script executes the ffmpeg binary using subprocess.run with an argument list. This is a secure practice that prevents shell injection vulnerabilities, even if filenames or AI-provided indices contain malicious characters.
[DATA_EXPOSURE] (SAFE): The script only accesses the video and subtitle files specified by the user. It writes the output segments to the same directory as the source video. No network activity or hardcoded credentials were detected.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill's workflow involves the AI reading and analyzing the content of subtitle files (VTT/SRT) to determine semantic split points. This creates a surface for indirect prompt injection where instructions embedded in subtitles could attempt to manipulate the AI's output.
Ingestion points: Subtitle files read in scripts/clip_video.py and processed by the AI.
Boundary markers: None; the AI is instructed to read the "full subtitle content".
Capability inventory: Subprocess execution of FFmpeg via subprocess.run.
Sanitization: None; the script parses the timestamps and text directly as provided in the file.

semantic-video-clipper