skills/naveedharri/benai-skills/video/Gen Agent Trust Hub

video

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill extensively utilizes CLI tools including ffmpeg, ffprobe, and remotion to process video and audio data. These operations are core to the skill's functionality and are performed on local files within the project directory.
  • [EXTERNAL_DOWNLOADS]: The implementation scripts include functionality to download the whisper.cpp binary and associated speech-to-text models. These downloads are managed by the @remotion/install-whisper-cpp library and target official repositories from the Remotion organization.
  • [INDIRECT_PROMPT_INJECTION]: The skill performs transcription of user-provided video audio and uses the text to identify cut points and highlights. While this ingests external content into the agent's context, the skill establishes clear logic for validating segments (e.g., sentence boundaries) and provides a QA testing framework to ensure output integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 03:09 AM
Security Audit — agent-trust-hub — video