Skills
skills/jianshuo/claude-skills/wjs-reframing-video/Gen Agent Trust Hub

wjs-reframing-video

Pass

Audited by Gen Agent Trust Hub on May 30, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads a required FaceLandmarker model file (~4 MB) from Google's official storage (storage.googleapis.com) during the first execution. This file is cached locally in the skill's model directory for subsequent use.
  • [COMMAND_EXECUTION]: The implementation uses the subprocess module to interact with ffmpeg and ffprobe for video probing and rendering. These operations are essential for the primary function of reframing video files and use structured command lists to prevent injection vulnerabilities.
  • [NO_CODE]: No evidence of malicious code, data exfiltration, or prompt injection was found during the analysis of the scripts and instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 30, 2026, 12:47 AM
Security Audit — agent-trust-hub — wjs-reframing-video