ffmpeg
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous shell command templates for
ffmpegandffprobe. These are standard tools for media processing and the commands provided (e.g., resizing, compression, trimming, concatenation) are consistent with the skill's stated purpose. - [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted media files and filenames.
- Ingestion points: Input file paths provided to
ffmpeg -iandffprobecommands inSKILL.mdandreference.md. - Boundary markers: None explicitly defined in the templates; the skill assumes the agent will handle shell argument escaping.
- Capability inventory: File system read/write via
ffmpegand shell redirection (e.g.,echo "file..." > list.txt). - Sanitization: Not present in the documentation templates, as they are examples for the agent to follow.
Audit Metadata