skills/zju-real/easel/audio-editing/Gen Agent Trust Hub

audio-editing

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs audio operations by executing a local Python script skills/shared/scripts/audio_ops.py, which acts as a wrapper for FFmpeg and FFprobe. The use of a dedicated script to manage command construction is a recommended practice to minimize command injection risks.- [INDIRECT_PROMPT_INJECTION]: The skill processes external media files, which represents an ingestion point for untrusted data.
  • Ingestion points: Audio and video files provided via the input_file parameter in SKILL.md.
  • Boundary markers: The instructions do not specify explicit delimiters for file content, but the data is processed by specialized binary tools (FFmpeg).
  • Capability inventory: The skill has the capability to run subprocesses (FFmpeg/FFprobe) and write to the filesystem.
  • Sanitization: The skill documentation explicitly states that parameters are determined by the script and commands are not manually spliced, which serves as a mitigation against injection.- [EXTERNAL_DOWNLOADS]: The skill references the official FFmpeg project website (ffmpeg.org) for software requirements and installation guidance. This is a well-known and trusted source for audio/video processing software.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 01:22 PM