music-to-video

Warn

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill contains an instruction for the agent to execute a self-update command (npx hyperframes skills update music-to-video) with the directive to "run silently, don't ask." This constitutes a concealment pattern that encourages the agent to bypass user confirmation for operations involving network access and code execution.
  • [COMMAND_EXECUTION]: The skill workflow relies on the execution of several system commands and scripts. The analyze-beatgrid.py script uses Python's subprocess module to call ffmpeg for audio conversion. Additionally, the instructions frequently employ the npx hyperframes CLI for project lifecycle management, and Node.js scripts like assemble-index.mjs and stage-assets.mjs perform direct file system operations including file copying and structured data generation.
  • [EXTERNAL_DOWNLOADS]: The skill initiates several external network requests for code and libraries. The npx update mechanism and package initialization fetch content from the npm registry. Furthermore, various components and generated output files reference external JavaScript libraries such as GSAP (v3.14.2 and v3.15.0) and Three.js (v0.147.0) via the cdn.jsdelivr.net Content Delivery Network. The skill also explicitly directs the installation of Python packages (librosa, numpy, soundfile) from PyPI.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 10, 2026, 02:18 AM
Security Audit — agent-trust-hub — music-to-video