render-editorial-motion-podcast
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides the agent with specific bash command templates for complex video and audio processing tasks using ffmpeg. Evidence is found in PIPELINE.md and SKILL.md which contain command strings for zoompan filters, audio clipping, and stream muxing.
- [INDIRECT_PROMPT_INJECTION]: The workflow involves processing user-provided audio files which are transcribed into text, creating a surface for potential indirect injection. Ingestion points include local podcast MP3 files and Whisper-generated JSON transcription data (words.json). There are no specific boundary markers or safety instructions to distinguish between data and instructions. The capability inventory includes the execution of shell commands and Python scripts. No explicit sanitization of transcribed text is described before it is used in captioning.
- [NO_CODE]: The skill payload consists entirely of documentation, instructions, and configuration examples. No executable scripts or binaries are included in the 6 files analyzed.
Audit Metadata