packaging-lead
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive analysis of the skill's SKILL.md file confirms it contains legitimate instructional content for video production workflows. No signs of prompt injection, obfuscation, or data exfiltration were found.
- [COMMAND_EXECUTION]: The skill references the use of
ffmpegfor extracting frames from video files. This is a standard and appropriate utility for a packaging lead role to generate thumbnail images from source video, and it is not used in a malicious or unauthorized manner. - [INDIRECT_PROMPT_INJECTION]: The skill processes project briefs and video transcripts as input. Ingestion points: 'Package Brief' section and video transcripts. Boundary markers: uses template headers but lacks explicit instructions to ignore embedded commands. Capability inventory: execution of
ffmpegand invocation of internal image generation tools. Sanitization: not specified. This constitutes an attack surface typical of content processing agents, but the absence of high-risk capabilities (like network access) renders the risk minimal.
Audit Metadata