ffmpeg-video-editor
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve interpolating untrusted user data into shell command templates, creating a vulnerability surface for command injection.
- Ingestion points: User-provided inputs such as filenames, filter parameters, and timestamps in
SKILL.md. - Boundary markers: While templates use double quotes (e.g.,
"INPUT","OUTPUT"), these do not prevent shell interpolation via backticks (`) or subshells ($()) in many terminal environments. No explicit boundary markers or instructions to ignore embedded commands are present. - Capability inventory: The skill generates commands for
ffmpeg, typically used with filesystem access and potentially network access (depending on FFmpeg build/protocols). - Sanitization: The instructions lack guidance on sanitizing or validating user-provided strings before they are embedded in executable shell commands.
- [COMMAND_EXECUTION]: The skill's primary function is the generation of shell commands for the
ffmpegutility. It specifically includes the-safe 0flag in the concatenation command template, which explicitly disables FFmpeg's internal filename safety checks, potentially allowing access to paths or protocols outside the intended scope if the input file list is manipulated.
Audit Metadata