claude-video-caption
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, prompt injections, or exfiltration attempts were detected. All described behaviors align with the stated purpose of video captioning.
- [COMMAND_EXECUTION]: The skill provides various commands for FFmpeg and Whisper-ctranslate2. These are standard multimedia tools, and their usage here is consistent with video processing workflows.
- [DYNAMIC_EXECUTION]: The skill includes an inline Python script for converting JSON timing data into ASS subtitle format. The script logic is transparently defined in the instructions and performs safe data transformation without external network or unauthorized file system access.
Audit Metadata