music-to-video
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection via user-supplied filenames and directory paths.
- Ingestion points:
scripts/analyze-beatgrid.py(audio input path) andscripts/stage-assets.mjs(source directory path via --from). - Boundary markers: Absent for file names or paths.
- Capability inventory:
subprocess.run(executes ffmpeg inscripts/analyze-beatgrid.py),writeFileSync(scripts/assemble-index.mjs), andcopyFileSync(scripts/stage-assets.mjs). - Sanitization: Absent; input paths are passed to shell tools and file APIs. This behavior is standard for local media processing utilities and is considered acceptable for the primary skill purpose.
- [EXTERNAL_DOWNLOADS]: The skill and its generated outputs reference legitimate animation and 3D libraries (GSAP, Three.js) hosted on trusted CDNs (jsDelivr) and fonts from Google Fonts. These references are standard for web-based animation and do not pose a security threat.
- [COMMAND_EXECUTION]: The skill executes local tools such as
ffmpegand thehyperframesCLI. These executions use safe parameter-passing methods (argument lists rather than shell strings) and are restricted to the primary task of media conversion and project management.
Audit Metadata